.;J,, .,,......_.,. •....

-• ·�-
.. ,c.,-

30 NEW DIRECTIONS FOR ELEMENTAR

Y SCHOOL MATHEMATICS

mals. When he gave a number

Sam a chance to answer a question on deci opened her mouth to ask a
3
she
response, the teacher's head came up and
B fore she cou ld spea k, Sam said, "I know, I know, Ms.
follow-up question. e
and to draw a picture! " He then
Davis. You want me to tell you why
proceeded to do exactly that.
ons and beliefs about mathemat- Developing Understanding in
It is not easy to change students' percepti
tions about what students can
ics. It is also not easy to change our own percep elp our students to value
and cannot do in mathematics. If we wan
t
er,
to h
and to have the confidence to
Mathematics via Problem Solving
mathematics, to develop mathematical pow
resting, challenging problem situa­
tackle new situations, we must pose inte
re, to formu late problems, to
tions and give our students time to explo
reason about the validity of these
develop strategies, to make conjectures, to Thomas L. Schroeder
ict, and, of course, to raise more
conjectures, to discuss, to argue, to pred
ents, showing them that we value Frank K. Lester, Jr.
questions! If we listen carefully to our stud
ren are remarkably clever at
their thoughts, we are likely to learn that child
making sense of mathematical situations.
REFERENCES
me­
rning and Fun with Geometry Games." Arith
Bright, George W., and John G. Harvey. "Lea
tic Teacher 35 (April 1988): 22-26.

E
ns from Grades 3 through 6. San Francisco
: Math
Burns, Marilyn. A Collection of MMh Lesso ARLY in th_is decad the theme of school mathematics shifted from "back
e
Solution Publications, 1987. ematics to the basics" to "problem solving." In fact, in recent years probl m
p Discussion?" Educational Studies in Math
e
Hoyles, Celia. ·'What Is the Point of Grou solving has been the most written-about and talked-about part of the math­
16 (2) (1985): 205-14. School
ics. Curriculum and Evaluation Standards for ematics curriculum and at the same time the least understood. Now that
National Council of Teachers of Mathemat The Council, 1987.
Math enwrics. Working draft. Reston, Va.: rencej,:
there has been nearly a decade of attempts to make problem solving "the
iculum for Mathematics." In Individual Diffe focus of school mathematics" (NCTM 1980, p. 1), we need to assess the
Romberg, Thomas A." A Common Curr G.D. Fenstermacher andJ. l. Goodlad, pp. 121-59.
and 1/u: Common Curriculum. edited by results of these efforts. This article addre sses the role of problem solving in
Chicago: Rand McNally & Co., I 976. rtu­ ,, elementary school mathematics in the hope of adding some much-needed
e-making: An Informal Attack on the Unfoe of
Schoenfeld, Alan H. •·on Mathematics as Sens emat ics." Pape r prese nted at the Offic ·,' clarity to the discussion. Our main point is that the most important role for
nate Divorce of Formal and Informal Math earning Research and Development Center
Con­
Educational Research and Improvement/L'i:ttion, University of Pittsburgh, March 1987. .;,·problem solving is to develop students' understanding of mathematics.
ference on Informal Reasoning and Educ
SbeU:
p. 13. Draft version. Nottingham, England:
Swan, Malcolm. The Langiwge of Graphs, ersity of Nottingham, 1986.
Centre for Mathematics Education, Univ ce;,; APPROACHES TO PROBLEM-SOLVING INSTRUCTION
nville: A Case Study." In Case Studies in Scien
Welch, William. "Science Education in Urba y, p. 6. Urbana, Ill.: University of Illinois, 19'.�u
F;ducntion, edited by R. Stak e and J. Easle i', -In the main, the discussions about problem solving and the efforts to
tevelop curricula and materials for students and teachers have been worth­
hile and helpful. Today the notion that problem solving should play a
l}ominent role in the curriculum has widespread acceptance. During the
··�it decade quite a large number of problem-solving resources have been
�veloped for classroom use in the form of collections of problems, lists of
'ategies to be taught, suggestions for activities, and guidelines for evaluat­
':. problem-solving performance. Much of this material has been very
)rul in helping teachers make problem solving a focus of their instruction.
,Wever, it has not provided tile sort of coherence and clear direction that is

31
 33
32 NEW DIRECTJONS FOR ELEMENTARY SCHOOL MATHEMATJCS DEVELOPING UNDERSTANDING VIA PROBLEM SOLVING

needed, primarily because to date little agreement has been reached on how Teaching via Problem Solving
as a
this goal is to be achieved. Undoubtedly there are several reasons for this In teaching via problem solving, problems are valued not only
state of affairs, but the confusion probably stems from the vast differences atics but also as a primar y means of doing so.
purpose for learning mathem
among individuals' and groups' conceptions of what it means to make matica l topic begins with a proble m situatio n that
The teaching of a mathe
ues are de­
problem solving the focus of school mathematics. One of the best ways of embodies key aspects of the topic, and mathematical techniq
coming to grips with these differences is to distinguish among three ap­ veloped as reasonable respon ses to reason able proble ms. A goal oflearning
proaches to problem-solving instruction: (1) teaching about problem solv­ mathematics is to transform certain nonroutine problems into routine ones.
ing, (2) teaching for problem solving, and (3) teaching via problem solving. The learning of mathematics in this way can be viewed as a movement from
the mathe­
An explicit statement of this distinction appeared in a paper written more the concrete (a real-world problem that serves as an instance of
ntation of a
than a decade ago by Hatfield (1978), but we suspect that others may have matical concept or technique) to the abstract (a symbolic represe
symbo ls).
espoused a similar point of view as well. Let us explain what each of these class of problems and techniques for operating with these
to
three approaches entails. sobre, para, através An example from the Middle Grades Mathematics Project can serve
and Fitzgerald 1986). A
illustrate teaching via problem solving (Shroyer
Teaching about Problem Solving the concep ts of area and
fifth-grade teacher who has decided to introduce
that
The teacher who teaches about problem solving highlights Polya's perimeter gives each student a set of twenty-four one-inch-square tiles
(1957) model of problem solving (or some minor variation of it). Briefly, this udents are challen ged to determ ine
are to be regarded as small tables. The st
banqu et tables of differe nt
model describes a set of four interdependent phases in the process of solving the number of small tables (tiles) needed to make
can be seated at these banqu et
mathematics problems: understanding the problem, devising a plan, carry­ sizes (area) and the number of people who
told t hat one small table can seat four
ing out the plan, and looking back. Students are explicitly taught the phases tables (perimeter). The students are
the small
that, according to Polya, expert problem solvers use when solving mathemat­ people, one on each side, and-that the banquet tables made from
The real-w orld situatio n (formi ng banquet
ics problems, and they are encouraged to become aware of their own tables are usually rectangular.
them) serves as a contex t in which studen ts
progression through these phases when they themselves solve problems. tables and seating people around
and the relatio nships betwee n them. At first no
Additionally, they are taught a number of "heuristics," or "strategies," from explore area and perimeter
les
which they can choose or which they should use in devising and carrying out formulas are used or developed; they will come in a later activity. Examp
ted by the teache r include the followi ng (Shroy er
their problem-solving plans. Some of the strategies typically taught include of the challenges presen
looking for patterns, solving a simpler problem, and working backward. At and Fitzgerald 1986):
its best, teaching about problem solving also includes experiences with Example A: Use your tiles to make different arrangements that will seat
actually solving problems, but it always involves a great deal of explicit twenty people.
discussion of, and teaching about, how problems are solved.
Example B: Add squares to the following arrangement so that the perime­
ter is 18. What is the new area?
Teaching for Problem Solving
In teaching for problem solving, the teacher concentrates on ways in

rrfjJ
which the mathematics being taught can be applied in the solution of both
routine and nonroutine problems. Although the acquisition of mathematical
· knowledge is of primary importance, the essential purpose for learning
mathematics is to be able to use it. Consequently, students are given many
instances of the mathematical concepts and structures they are studying and
many opportunities to apply that mathematics in solving problems. Further, Some Observations about the Three Approaches
the teacher who teaches for problem solving is very concerned about stu­ Although in theory these three conceptions of teaching problem solving in
dents' ability to transfer what they have learned from one problem context to mathematics can be isolated, in practice they overlap and occur in various
others. A strong adherent of this approach might argue that the sole reason combinations and sequences. Thus, it is probably counterproductive to
for learning mathematics is to be able to use the knowledge gained to solve argue in favor of one or more of these types of teaching or against the others.
problems.
 35
34 NEW DIRECTIONS FOR ELEMENTARY SCHOOL MATHEMATICS DEVELOPING UNDERSTANDING VIA PROBLEM SOLVING

Nevertheless, if curriculum developers, textbook writers, or classroom TWO MODElS OF THE PROCESS OF SOLVING
teachers intend to make problem solving the "focus of instruction," they MATHEMATICS PROBLEMS
need to be aware of the limitations inherent in exclusive adherence to either Problem solving has sometimes been conceptualized in a simplistic way by
of the first two types of problem-solving instruction. One such limitation a model like that in figure 3.1. This model has two levels, or "worlds": the
stems from the fact that problem solving is not a topic of mathematics, and it everyday world of things, problems, and applications of mathematics; and
should not be regarded as such. If teaching about problem solving is the the idealized, abstract world of mathematical symbols, operations, and
focus, the danger is that "problem solving" will be regarded as a strand to be techniques. In this model the problem-solving process has three steps:
added to the curriculum. Instead of problem solving serving as a context in Beginning with a problem posed in terms of the everyday physical reality,
which mathematics is learned and applied, it may become just another topic, the problem solver first translates (arrow A) the problem into abstract
taught in isolation from the content and relationships of mathematics. mathematical terms, then operates (arrow B) on the mathematical repre­
A different shortcoming arises from teaching for problem solving. When sentation to come to a mathematical solution of the problem, which is then
this approach is interpreted narrowly, problem solving is viewed as an translated back (arrow C) into the terms of the original problem.
activity students engage in only after the introduction of a new concept or
following work on a computational skill or algorithm. The purpose is to give Mathematical World
students an opportunity to "apply" recently learned concepts and skills to
the solution of real-world problems. Often these problems appear under a
heading such as "Using Division to Solve Problems," and a solution of a
sample story problem is given as a model for solving other, very similar
problems. Often, solutions to these problems can be obtained simply by
following the pattern established in the sample, and when students en­ Seguir o exemplo
counter problems that do not follow the sample, they often feel at a loss. It Ensino PARA RP
has been our experience (which is supported by several studies) that when
taught in this way, students often simply pick out the numbers in each
problem and apply the given operation(s) to them without regard for the "Real" World
problem's context; as often as not, they obtain the correct answers. In our
view this practice is certainly not problem solving. Indeed, it does not even Fig. 3.1. A simplistic model of the process of solving mathematics problems
require mathematical thinking. Furthermore, a side effect is that students According to this model, mathematics can be, and often is, learned sepa­
come to believe that all mathematics problems can be solved quickly and
rately from its applications. In teaching for problem solving, instructors are
relatively effortlessly without any need to understand how the mathematics very concerned to develop students' abilities to translate real-world prob­
they are using relates to real situations. Unfortunately, this approach to lems into mathematical representations, and vice versa. But they tend to
problem-solving instruction has been quite common in textbooks. deal with problems and applications of mathematics only after those math­
Unlike the other two approaches, teaching via problem solving is a ematical concepts and skills have been introduced, developed, and prac­
conception that has not been adopted either implicitly or explicitly by many ticed. The difficulty with this model is that it applies to routine problems
teachers, textbook writers, and curriculum developers, but it is an approach better than to nonroutine ones. Problems classified as "translation prob­
to the teaching of mathematics that deserves to be considered, developed, lems" (Charles and Lester 1982) are solved exactly as the model indicates,
tried, and evaluated. Indeed, teaching via problem solving is the approach but for more challenging problems, like those categorized by Charles and
that is most consistent with the recommendations of NCTM's Standards Lester as "process problems," the problem solver has no single already­
Commission that (1) mathematics concepts and skills be learned in tile learned mathematical operation that will solve the problem. As well as
context of solving problems; (2) the development of higher-level thinking translation and interpretation, these nonroutine problems also demand
processes be fostered through problem-solving experiences; and more complex processes, such as planning, selecting a strategy, identifying
(3) mathematics instruction take place in an inquiry-oriented, problem­ subgoals, conjecturing, and verifying that a solution has been found. For
solving atmosphere (NCTM 1987). nonroutine problems, a different type of model is required.
36 NEW DIRECTIONS FOR ELEMENTARY SCHOOL MATHEMATICS DEVELOPING UNDERSTANDING VIA PROBLEM SOLVING 37
Figure 3 .2 shows a modification of the problem-solving model for transla­ the mathematical world and proceed directly along arrow X' to a mathemat­
tion problems that can be used to illustrate thinking processes when nonrou­ ical generalization and hence to a solution of the original real-world prob­
tine problems are involved and when teaching via problem solving is lem. In such a situation the solution process can be modeled as shown in
adopted. This model also contains two levels that represent the everyday figure 3.1.
world of problems and the abstract world of mathematical symbols and
operations. In this model, however, the mathematical processes in the upper PROBLEM SOLVING AND UNDERSTANDING IN MATHEMATICS
level are "under construction" (i.e., being learned, as opposed to already
learned), and its most important features are the relationships between the Central to our interest in teaching via problem solving is the belief that the
steps in the mathematical process (in the upper level) and the actions on primary reason for school mathematics instruction is to help students under­
particular elements in the problems (in the lower level). stand mathematical concepts, processes, and techniques. During the back­
to-basics movement of the 1970s, and also with the more recent focus on
Mathematical World problem solving, this fundamental tenet of good mathematics instruction has
been given far too little attention. Moreover, some commentators have
limited their discussion of understanding to the question of students' com­
ry!,ithe[iD�ji¢$f. X' prehension of the information presented in mathematical text, especially in
. r�prese.cts-tioD
the statements of verbal problems. In our view, students' understanding of
mathematics involves much more than this.
Y, A large number of mathematics educators have written about mathemati­
cal understanding by distinguishing between types or qualities of under­
standing. Brownell's work (e.g., 1935, 1945, 1947) on "meaningful arithme­
tic" in the 1930s and 1940s is especially relevant, but only during the past ten
X
to fifteen years has any substantial activity taken place in this area. Of
particular note are the works of Skemp (1976, 1979), Herscovics and
Bergeron (1981, 1982), Davis (1984), and Hiebert (1984, 1986). A com­
"Real" World mon thread running through these considerations of the nature of under­
Fig, 3.2. A model of the process of solving process problems
standing in mathematics is the idea that to understand is essentially to relate.
In particular, a person's understanding increases (1) as he or she is able to
In the figure, some of the Y arrows point upward to indicate that the relate a given mathematical idea to a greater number or variety of contexts,
problem solver is learning to make abstract written records of the actions or (2) as he or she relates a given problem to a greater number of the
mathematical ideas implicit in it, or (3) as he or she constructs relationships
that are understood in a concrete setting. These arrows pointing upward
among the various mathematical ideas embedded in a problem.
represent the processes of abstraction and generalization. Some of the
Indications that a student understands (or misunderstands, or does not
arrows point downward to show that the problem solver is able to explain a
understand) specific mathematical ideas often appear as the student solves a
mathematical process by referring to the real-world actions that the mathe­
problem. Relationships of the kinds mentioned above are evident in stu­
matical symbols represent. Arrows pointing downward might also suggest
dents' attempts to solve the following problem (fig. 3.3), which is an adapta-
.that a problem solver who had forgotten the details of a mathematical
procedure would be able to reconstruct that process by imagining the corre­
sponding concrete steps in the world in which the problem was posed. The �@@��€i�@
collection of Y arrows illustrate the correspondence between the process of
solving the problem in concrete terms (labeled X) and the parallel, abstract �®€i@@)©©®
mathematical process (labeled X'). The Y arrows also show that the prob­ @®�@@®®@
lem solver typically moves back and forth between the two worlds-the real
and the mathematical-as the need arises. For a particular problem the
®@®�
problem solver might move directly along arrow Y 1 from the real world to Fig. 3.3. A coin problem
38 NEW DIRECTIONS FOR ELEMENTARY SCHOOL MATHEMATICS DEVELOPING UNDERSTANDING VIA PROBLEM SOLVING 39
tion of one suggested for students in grade 2 (Alberta Education 1983, p.
54). This problem was used by Croft (1987), a teacher who conducted @@ti@ 08®®
individual interviews with several of her grade 1 students. @@@) ®@ �©�®
Croft noticed that different children had distinct levels of understanding
that corresponded to the number of different mathematical concepts and ®@@®
processes they used in solving the problem. All the children began by sorting
the coins by value and repeatedly using one-to-one correspondence; they
placed the four dimes into four piles, then four of the nickels, then four more �«w®
nickels, then the pennies. But some students got stuck when they reached the Fig. 3.5. Some different ways to make twenty-five cents
situation shown in figure 3.4. Although they were satisfied that the coins in
each of the four piles matched, they did not know what to do with the nickel
ability to recognize and use these ideas gives a measure of their understand­
and three pennies that were left over. The children at the lowest level of
ing. It is interesting to note that some children's understanding seemed to
understanding never got beyond this impasse, despite being encouraged to
deepen and grow as they worked on the problem; their progress with the
"try a different way" and to "share out all the coins." One child suggested
problem came in stages, by discovery, rather than all at once. This suggests
that he could solve the problem if only he could take the nickel to the store
that these students were learning via problem solving, even though the
and exchange it for five pennies. Other children used the making-change
teacher's purpose was to assess their understanding rather than to teach
idea and recognized that by "undoing" part of the sharing, they could
them via problem solving.
remove the three pennies from one pile and replace them with the "extra"
nickel so that six pennies could then be put in the other three piles. This
solution represents the next level of understanding. When the teacher asked, DEVELOPING UNDERSTANDING VIA PROBLEM SOLVING
"Can you find another way? Could the piles have different coins in them but
the same amount of money?" some children rearranged the coins but found We believe that instead of making problem solving the focus of mathemat­
no new solutions. However, the students at the highest level of understand­ ics instruction, teachers, textbook authors, curriculum developers, and eval­
ing noticed that the value of the coins in each pile was twenty-five cents and uators should make understanding their focus and their goal. By doing so
used this fact to find several different ways of making change, such as those they will shift from the narrow view that mathematics is simply a tool for
shown in figure 3.5. solving problems to the broader conception that mathematics is a way of
These differences in students' performance indicate the variety of mathe­ thinking about and organizing one's experiences. As a consequence, prob­
matical operations inherent in the problem, including sorting, finding a lem solving will not be de-emphasized, but the role of problem solving in the
one-to-one correspondence, iterating, exchanging sets of equal value, and curriculum will change from being an activity students engage in after they
counting the value of a collection of coins and using the value (rather than have acquired certain concepts and skills to being both a means for acquiring
the coins themselves) to find other collections of the same value. The pupils' new mathematical knowledge and a process for applying what has been
learned previously. Fundamental to the view that understanding should be a
«w ,•'' � 0 �
primary goal of instruction is the belief that children's learning of mathemat­

®@ ti� ti®
\
8® ics is richest when it is self-generated rather than when it is imposed by a
teacher or textbook. A primary advantage of self-generated knowledge is
that it is tied to what the learner already knows. Furthermore, when children
@® @'J@ @� �� construct new mathematical knowledge for themselves, they learn not only
@) @ ® @) concepts, facts, skills, and so on, but also how to manage and regulate the
application of this new knowledge. That is, they are in charge of this
knowledge (and of their learning in general), thereby making it more useful
8©@'J@ to them in solving problems and in learning new concepts and skills. A
benefit of having acquired mathematical knowledge in this way is that
Fig. 3.4. An impasse on the way to solving the coin problem problem-solving efforts are less susceptible to error. We believe that teach-

Diferença do ensino para e através

40 NEW DJRECTIONS FOR ELEMENTARY SCHOOL MATHEMATICS DEVELOPING UNDERSTANDlNG VIA PROBLEM SOLVING 41

ing via problem solving and teaching for understanding are not only compat­ l. Understanding increases the richness of the types of representations that
ible but in fact mutually beneficial. the problem solver can construct. During problem solving it is necessary for
the problem solver to internalize the information in a problem. That is, the
Problem Solving Enhances Understanding problem solver must develop a representation of the information. The more
Beatriz D' Ambrosio has suggested the following challenge as an illustra­ accurately the representation depicts the information and links pieces of
tion of the fact that solving a problem can deepen a student's understanding information together, the more likely it is that the problem will be solved
of a topic of mathematics. correctly.
2. Understanding assists the problem solver in monitoring the selection and
On centimeter graph paper outline all the shapes that have an area of 14
execution of procedures (e.g., strategies, a lgorithms). Successful problem
square cm and a perimeter of 24 cm. For each shape you draw, at least
solving requires the ability to monitor the selection and subsequent execu­
one side Oi each square must share a side with another square. Here's an
tion of procedures, and the ability to evaluate the extent to which local
example:
actions (e.g., performing computations) conform to goals, and the ability to
make various trade-off decisions ( e.g., deciding that an estimate will give a
Allowed Not allowed "close enough" answer). The problem solver who understands the relation­

EP
ships among the conditions and variables in a problem and who can place the
problem in a meaningful context is well equipped to anticipate the conse­
quences of various decisions and actions and to evaluate the progress being
� made toward a solution.

It is assumed, of course, that students given this problem would already 3. Understanding aids the problem solver in judging the reasonableness of
results. The ability to create a meaningful and appropriate internal repre­
have a basic understanding of the concepts of area and perimeter for rectan­
sentation of the information in a problem enhances the problem solver's
gular shapes. The intention is not simply to allow the students an opportunity
ability to determine whether the answer makes sense.
to apply their knowledge of these two concepts. Rather, it is to enhance their
understanding of the relationships between area and perimeter. The solu­ 4. Understanding promotes the transfer of knowledge to related problems
tion to this problem requires that students make many decisions, among and its generalizability to other situations. Brownell (1947), among others,
them how to keep track in a systematic way of the shapes that have been has pointed out that a solution to a problem that is meaningful (i.e., well
made so that all possibilities will be found and none will be duplicated. Such understood) transfers readily to problems that are similar in structure even if
decisions and the associated skills needed to carry them out are an important they are different in context. That is, since understanding involves the ability
part of learning how to solve problems successfully and efficiently. But to apply a particular concept, skill, or procedure �o unfamiliar situations, an
learning what decisions to make and when to make them is not the only individual who has a good understanding of certain mathematical ideas and
benefit of this task. In addition, as shapes are modified to fit the conditions of techniques is likely to be able to apply that learning to contexts that might be
the problem, the learner is exposed to relationships between area and very different from the contexts in which the mathematics was originally
perimeter that, if noticed, can facilitate a richer understanding of both learned.
concepts. Thus, through investigation and exploration, students not only
learn some useful problem-solving skills but also deepen their understand­ CONCLUSION
ing of two important measurement concepts.
We believe that there can be a mutually supportive relationship between
Understanding Aids Problem Solving emphasizing problem solving and emphasizing understanding in mathemat­
Of course, success in solving a problem depends on the student's having a ics instruction. When teachers teach via problem solving, as well as about it
good understanding of the information in it. However, the value of under­ and for it, they provide their students with a powerful and important means
standing in successful problem solving goes far beyond this. In particular, of developing their own understanding. As students' understanding of
when understanding is viewed in the way we have discussed, it aids problem mathematics becomes deeper and richer, their ability to use mathematics to
solving in at least four distinct ways. solve problems increases.