Thinking skills in realistic mathematics

J.M.C. Nelissen

Introduction

One of the most enduring ideas concerning mathematics instruction is the following:
mathematics consists of a set of indisputable rules and knowledge; this knowledge has a fixed
structure and can be acquired by frequent repetition and memorization. In the past twenty-five
years, far-reaching changes have taken place in mathematics instruction. More than in any other
field, such changes were influenced by mathematicians who had come to view their discipline in
a different light. Their observations went a long way towards stimulating a process of renewal in
mathematics instruction. New consideration was given to such fundamental questions as: how
might mathematics best be taught, how might children be encouraged to show more interest for
mathematics, how do children actually learn mathematics, and what is the value of mathematics?
According to Goffree, Freudenthal, and Schoemaker (1981), the subject of mathematics is
itself an essential element in ‘thinking’ through didactical considerations in mathematics
instruction. Moreover, the notion is emphasized that knowledge is the result of a learner’s
activity and efforts, rather than of the more or less passive reception of information. Mathematics
is learned, so to say, on one’s own authority. From a teacher’s point of view there is a sharp
distinction made between teaching and training. To know mathematics is to know why one
operates in specific ways and not in others. This view on mathematics education is the basic
philosophy in this chapter (Von Glazersfeld, 1991) In order to understand current trends in
mathematics education, we must consider briefly the changing views on this subject.
The philosophy of science distinguishes three theories of knowledge. Confrey (1981) calls
these absolutism, progressive absolutism and conceptual change. In absolutism, the growth of
knowledge is seen as an accumulation, a cumulation of objective and empirically determined
factual material. According to progressive absolutism a new theory may correct, absorb, and
even surpass an older one. Proponents of the idea of conceptual change have defended the point
of view that the growth of knowledge is characterized by fundamental (paradigmatical) changes
and not by the attempt to discover absolute truths. One theory may have greater force and present
a more powerful argument than another, but there are no objective, ultimate criteria for deciding
that one theory is incontrovertibly more valid than another (Lakatos, 1976). Mathematics has
long been considered an absolutist science. According to Confrey (1981), it is seen as the
epitome of certainty, immutable truths and irrefutable methods. Once gained, mathematical
knowledge lasts unto eternity; it is discovered by bright scholars who never seem to disagree,
and once discovered, becomes part of the existing knowledge base.
Leading mathematicians however have now abandoned the static and absolutist theory of
mathematics (Whitney, 1985). Russell (in Bishop, 1988) once explained that mathematics is the
subject in which we never know what we are talking about, nor whether what we are saying is
true. Today mathematics is more likely to be seen as a fluctuating product of human activity and
not as a type of finished structure (Freudenthal, 1983). Mathematics instruction should reveal
how historical discoveries were made. It was not (and indeed is still not) the case that the
practice of mathematics consists of detecting an existing system, but rather of creating and
discovering new ones. This evolving theory of mathematics also led to new ideas concerning
mathematics instruction. If the essence of mathematics were irrefutable knowledge and ready-
made procedures, then the primary goal of education would naturally be that children mastered
this knowledge and these procedures as thoroughly as possible. In this view, the practice of
mathematics consists merely of carefully and correctly applying the acquired knowledge If,
however, mathematicians are seen as investigators and detectives, who analyse their own and
others’ work critically, who formulate hypotheses, and who are human and therefore fallible,
then mathematics instruction is placed in an entirely different light. Mathematics instruction
means more than acquainting children with mathematical content, but also teaching them how
mathematicians work, which methods they use and how they think. For this reason, children are
allowed to think for themselves and perform their own detective work, are allowed to make
errors because they can learn by their mistakes, are allowed to develop their own approach, and
learn how to defend it but also to improve it whenever necessary. This all means that students
learn to think about their own mathematical thinking, their strategies, their mental operations and
their solutions.
 Mathematics is often seen as a school subject concerned exclusively with abstract and formal
knowledge. According to this view, mathematical abstractions must be taught by making them
more concrete. This view has been opposed by Freudenthal (1983) among others. In his opinion,
we discover mathematics by observing the concrete phenomena all around us. That is why we
should base teaching on the concrete phenomena in a world familiar to children. These
phenomena require the use of certain classification techniques, such as diagrams and models (for
example, the number line or the abacus). We should therefore avoid confronting children with
formal mathematical formulas which will only serve to discourage them, but rather base
instruction on rich mathematical structures, as Freudenthal calls them, which the child will be
able to recognize from its own environment. In this way mathematics becomes meaningful for
children and also makes clear that children learn mathematics not by training formulas but by
reflecting on their own experiences.
In the 1970s, the new view of mathematics, often referred to as mathematics as human
activity, led to the rise of a new theory of mathematics instruction, usually given the designation:
realistic. As it now appears, this theory is promising, but it is not the only theoretical approach in
mathematics instruction; three others can be distinguished: the mechanistic, the structuralist and
the empirical (Treffers, 1991). In the following we just give a brief characteristic of each
approach, because it is beyond the scope of this chapter to discuss the three schools in extension:
− The mechanistic approach reflects many of the principles of the behaviouristic theory of
learning; the use of repetition, exercises, mnemonics, and association comes to mind. The
teacher plays a strong, central role and interaction is not seen as an essential element of the
learning process. On the contrary, mathematics class focuses on conclusive standard
procedure.
− According to the second approach − the structuralist − thinking is not based on the children’s
experiences or on contexts, but rather on given mathematical structures. The structuralist
tends to emphasize strongly the teacher’s role in the process of learning.
− The outstanding feature of the third trend − the empirical − is the idea that instruction should
relate to a child’s experiences and interests. Instruction must be child-oriented. Empiricists
believe that environmental factors form the most important impetus for cognitive
development (Papert, 1980). Empiricists emphasize spontaneous actions.
Realistic mathematics instruction as progressive mathematization

In this section we present five features which characterize realistic mathematics. At first we are
dealing with learning in a context and second with the use of models. The third point (the
mathematical subjects are not atomized but interwoven) is not of so much relevance for this book,
while the three characteristics of the process of mathematization (construction, reflection, and
interaction) are analysed in the following sections.
The new realistic approach to learning and thought process in children has far-reaching
consequences. Mathematization is viewed as a constructive, interactive and reflective activity.
To begin, the point of departure for education is not learning rules and formulas, but rather
working with contexts. A context is a situation which appeals to children and which they can
recognize in theory. This situation might be either fictional or real, and forces children to call
upon the knowledge they have gained by experience − for example in the form of their own
informal working methods − thereby making learning a meaningful activity for them, A well-
chosen context can induce an active thought process in children, as the following example shows.
Let us start to give children of, say, 11 years the following formal and bare problem, not
presented in a context: 6 ÷ 3 4 . Many of them will have a great deal of trouble finding a solution
(Streefland, 1991). Some will answer, for example: 2 4 , 3 24 or 4 1 2 . They manipulate at
random with the given numbers, for instance 6 ÷ 3 = 2, so 6 ÷ 3 4 must be 2 4 . This child views
fractions as whole numbers and so do other students (Lesh et al., 1987). But some students will
calculate that 6 × 4 = 24 and that 24 divided by 3 equals 8. It is true that the latter answer is
correct, but when these children are questioned more closely, it turns out that they understand
almost nothing about the operation which they themselves have just performed. They just
remembered a rule they learned by heart, they know that the given solution is correct however
they don’t know why.
Now, the same children are next given the following context problem which is accompanied
by a picture: a patio is 6 metres long; you want to put down new bricks and the bricks you are
going to use measure 75 centimetres in length ( 3 4 of a metre). How many bricks will you need
for the length? This problem is the same as the previous one, but it has now been presented
within a context, a picture of a patio and the bricks to put down. This presentation elicits a
child’s own, informal approach: measuring out. This approach provides insight into the problem,
something which the symbolic form (6 ÷ 3 4 ) did not do. Some students even manipulated and
took the measure in reality, this means they measured out step by step 75 centimetres and after 8
steps they counted 6 metres. So the answer must be ‘eight’, they concluded. This example
demonstrates that working with contexts − which, if carefully constructed, can be considered
paradigmatic examples − form the basis for subsequent abstractions and for conceptualization.
That is because thinking must achieve a higher, abstract level and at that level these particular
contexts no longer serve a purpose. That is not to say that a process of decontextualisation occurs,
but rather recontextualisation. The children continue to work with contexts, but these contexts
become increasingly formal in nature; they become mathematical contexts. Their connection
with the original context, however, remains clear. The process by which mathematical thinking
becomes increasingly formal is called the process of progressive mathematization. Contexts, thus,
have various functions. They may refer to all kind of situations and to fantasy situations (Van
den Heuvel-Panhuizen, 1996). It is important that the context offer support for motivation as well
as reflection. A context should indicate certain relevant actions (to take measures in the example
above), provide information which can be used to find a solution-strategy and/or a thinking-
model.
Of course, leaving the construction to the students does not guarantee the development of
successful strategies. However it guarantees that students get the opportunity to practice
mathematician’s thinking and problem solving processes. Strategies are tried, tested and
elaborated in various situations.
In the previous discussion we have not argued that a student presented with ‘bare’ numerical
tasks (like 6 ÷ 3 4 ) will necessarily fail to solve the problem. Hence we were not suggesting
either that students who are given context problems will necessarily produce the right solution.
In recent research there is found a strong tendency of children to react to context problems
(‘word problems’) with disregard for the reality of the situations of these problems. Let us give
two examples of items used in research (Greer, 1997; Verschaffel et al., 1997):
− ‘An athlete’s best time to run a mile is 4 minutes and 7 seconds. About how long would it
take him to run 3 miles?’
− ‘Steve has bought 4 planks of 2.5 metre each. How many planks of 1 metre can he get out of
these planks?’
 In four studies, discussed by Greer (1997), the percentage of the number of students
demonstrating any indication of taking account of realistic constraints is: 6%, 2%, 0% and 3%.
The student’s predominating tendency to apply rules clearly formed an impediment to
thoroughly understanding the situation.
Verschaffel et al. (1997) confronted a group of 332 students (teachers in training) with word
problems and found they produced ‘realistic’ responses in only 48% of cases. Moreover the pre-
service teachers considered these ‘complex and tricky word problems’ as inappropriate for (fifth
grade) children. The goal of teaching word problem solving in elementary school, after their
opinion, was “...learning to find the correct numerical answer to such a problem by perforn1ing
the formal-arithmetic operation(s) ‘hidden’ in the problem” (Verschaffel et al., 1997, p. 357).
When solving word problems students should go beyond rote learning and mechanical
exercises to apply their knowledge (Wyndhamn & Säljö, 1997). Their research showed that
students (10-12 years of age) gave in most cases logically inconsistent answers. The authors
interprete these findings by claiming that the students focus on the syntax of the problem rather
than on the meaning. That means that the well-known rule-based relationship between symbols
results in less of attention being paid to the meaning. The students follow another ‘rationality’,
that is, they consider word problems as mathematical exercises “… in which a algorithm is
hidden and is supposed to be identified.” (Wyndhamm & Säljö, p. 366). Hence they do not know
or realize that they are expected to solve a real life problem.
Reusser and Stebler (1997) discuss another interesting research finding namely the fact that
pupils ‘solved’ unsolvable problems without ‘realistic reactions’. For example:
− ‘There are 125 sheep and 5 dogs in a flock. How old is the shephard?’ (Greer, 1997).
A pupil questioned by the investigators gave as his opinion: ‘It would never have crossed my
mind to ask whether this task can be solved at all’. And another pupil said: ‘Mathematical tasks
can always be solved’. One of the author’s conclusions is that a change is needed from
stereotyped and semantically poor, disguised equations to the design of intellectually more
challenging ‘thinking stories’. What we need are better problems and better contexts. Finally,
Reusser and Stebler (1997) − following Gravemeijer (1997) − give as their interpretation of the
research findings that the children are acting in accordance with a typical school mathematics
classroom culture.
 Second, the process of mathematization is characterized by the use of models. Some
examples are schemata, tables, diagrams, and visualizations. Searching for models − initially
simple ones − and working with them produces the first abstractions. Children furthermore learn
to apply reduction and schematization, leading to a higher level of formalization. We will
demonstrate, once again this using the previous example. To begin, children are able to solve the
brick problem by manipulating concrete materials. For instance, they might attempt to see how
often a strip of paper measuring 3 4 of a metre fits in a 6-metre-long space. At the schematic
level, they visualize the 6-metre-long patio and draw lines which mark out each 3 4 of a metre or
75 centimetres. The child adds 75 + 75 + 75... until the 6 metres have been filled The
visualization looks as follows:

----- / ----- / ----- / ----- / ----- / ----- / ----- / ----- /

34 34 34 34 34 34 34 34

An example of reasoning on a formal-symbolic level is as follows: 75 centimetres fits into 3

metres 4 times. We have 6 metres, so we need 2 × 4 = 8 bricks. The formula initially tested can
also be applied, but this time with insight: 1 4 metre fits 4 times into 1 metre, so it fits 24 times
into 6 metres. But I only have 3 4 of a metre, so I have to divide 24 by 3, and that makes 8. At
this formal level, moreover, the teacher can also explore the advantages and disadvantages of the
two methods with the children.
Third, an important element of realistic mathematics instruction is that subjects and curricula
(such as fractions, measurement and proportion) are interwoven and connected, whereas in the
past, the subject matter was divided − and so atomized.
Fourth, two other important characteristics of the process of mathematization are that it is
brought about both by a child’s own constructive action and by the child’s reflections upon this
action.
Finally, learning mathematics is not an individual, solitary activity, but rather an interactive
one.
Construction

Learning mathematics is a constructive activity, an aspect which has been emphasized by many
authors (including Bruner, 1986, 1996; Cobb, 1994; Cobb et al., 1997; Resnick & Klopfer, 1989;
Steffe, Cobb, & Von Glazersfield, 1988). Children construct internal, mental representations.
These might be concrete images, schemata, procedures, working methods at the abstract-
symbolic level, intuitions, contexts, schemata of solutions, or thought experiments. To make
clear how individuals construct different kinds of representations we present now a example (this
example is above the cognitive level of school children).
Suppose we were able to tighten a rope around the equator so that it is taught and lies flat on
the surface. We cut the rope and insert a metre-long piece of rope between the two cut ends.
Once again we tighten the rope so that it is taught all around. The question is: How far above the
ground is the rope now? The mental representations of many adults will contain various elements.
To start, they will ask themselves what the circumference of the earth is: 44,000 kilometres. The
rope − they can picture it before their very eyes − must therefore measure the same in length.
Another metre − they reason to themselves − scarcely matters in proportion to that enormous
distance. Probably the rope barely lifts off the ground. These mental representations actually
consist of concrete images which form a basis for solving the problem, conceived of as the
relationship between that one metre and the entire circumference of the earth. A mathematically
trained problem-solver will construct entirely different representations. He or she will
immediately dismiss any concrete facts and reduce the earth to a circle, focusing in on the
relationship between the radius and the circumference. This relationship is then converted into a
formula, 2π R. The representation is created by converting a concrete problem into an adequate
mathematical formula. The rope lies about 16 centimetres above the surface of the earth.
Learning mathematics as a constructive activity means that a child’s own discoveries are
taken seriously. This does not mean that their discoveries are always on the mark, but they do
give the teacher a recognizable handle from which he or she can begin to teach. The teacher
learns the general outlines of the representations of children and can adjust his approach
accordingly. But what is the function of representations in mathematics instruction? The
representational point of departure and the ‘representational view of mind’ seems to require
some constructivist comment. What is criticized and rejected is the metaphor of the mind as a
mirror that reflects a mathematically prestructured environment unaffected by individual and
collective human activity (Von Glazersfeld, 1991). Correct, internal representations are
constructed by confronting them with external representations. These are socially and culturally
determined (Cobb, 199). Children do indeed actively construct their own mathematical
knowledge, but their purpose is to participate increasingly in taken-as-shared mathematical
practices. These practices are played out both in the classroom as in society and science.
If rules and procedures are prescribed prematurely and one-sided, blocking a child’s own
representations, problems will ensue. The following recorded fragment of conversation serves as
a concrete illustration. Henry, a good pupil of 9 years of age, is busy working out subtraction
problems in his mathematics workbook. The book gives the following formula to complete the
problems:
94 - 52 = ... - ... - ... = ... - ... = ...
Researcher: How do you do that?
Henry: First you subtract 50, that’s 44, and then you subtract 2, making 42.
Researcher: Do you always have to do it that way?
Henry: Yes.
Researcher: Can’t you subtract 2 first?
Henry: That’s not allowed.
Researcher: But why not?
Henry: Because the book says. (He points out the following example: 54 -
31 = ., 54 - 30 - 1=., 24 – 1 = 23)
Researcher: What if I subtract 2 first anyway?
Henry: But that’s against the rules.
Researcher: Will the answer be different if you subtract 2 first?
Henry: Maybe.
Let us now look at an example in which children are given a change to develop their own
constructions. A group of 10 to 11-year old students was asked how they would go about solving
the following problem. There are a number of bottles on a table, and each bottle has a different
shape. None of them has labels, so no one scan tell just how much each bottle can hold. How
would you figure out which bottle can hold the most water? The children were asked to present
their ideas and talk about one another’s ideas. One child suggests weighing the bottles. No,
another says, hold them under water and see how much the water rises. A third suggests dumping
the contents on the floor and seeing which puddle is the biggest. This is good example of a
practical situation in which children are constructing knowledge, taken-as-shared (Cobb, 1994).
Note that here is qualified the word ‘shared’. The children’s solutions do not match precisely, but
they are considered ‘compatible’ and are therefore worth discussing. And this is what happened.
The children were criticizing and commenting each other solutions until one child proposed to
use a glass as a measure. All the children insisted with this idea but now rose the question of how
big the glass should be. At the end of the discussion they decided to choose a small glass but not
too small. The idea that children’s own constructions form the point of departure for the
teaching-learning process in mathematics instruction is one of the fundaments of the realistic
school. Confrey (1985) argues that a person’s knowledge is necessarily the product of his/her
own constructions or mental acts. Thus s/he can have no direct or unmediated knowledge of any
objective reality. Knowledge is created by means of images or representations and these are
products of our mental actions (Gardner, 1987). But if their own constructions are so very
important, children should be allowed to nurture their own constructions (whatever their quality
may be). It is not necessarily that this would lead to anarchy and blocked communication during
mathematics class, because constructions arise through interaction with other children and with
the teacher. Bruner (1986) too asserts that constructivism is not a sort of cultural relativism or an
homage to the proposition that ‘anything goes’. Neither should constructions be understood in
Piagetian terms. Piaget (1976) was concerned with individual constructions which arise from the
subject’s own position and which are the result of intrinsic and autonomous processes;
‘mathematical practices’ have relatively little influence on them.
Lo, Grayson, Wheatly, and Smith (1990) discuss the close relationship between construction
and interaction in the following fashion: “From a constructivist’s perspective learning occurs
when a child tries to adapt her functioning schemes to neutralize perturbations that arise through
interactions with our worlds” (p. 116). Two important aspects, constructions and interactions, are
important in the above statement. Although construction of knowledge is a personal act, it is by
no means an isolated activity as many people’s interpretations of constructivism imply.
Constructivists recognize the importance of social actions as ‘the most frequent source of
perturbations’ (Salomon, 1989). Interactions, thus, lead to construction, because the process of
interacting often causes perturbations in the normal pattern of behaviour − particularly when
unexpected problems arise − which the person involved will try to resolve by seeking his or her
own solutions (constructions). A study conducted by Saxe (1988) investigated how young
Brazilian candy sellers, who generally had little or no schooling, had learned mathematics. These
children had learned to fix cost prices, to calculate skilfully in cash amounts, and to think in
ratios (3 pieces of candy is 500 cruzeiros, 1 piece is 200), but when confronted with classroom
problems − for example, reading and comparing double-digit numbers − they were at a complete
loss. They had created their own constructions in the process of solving problems encountered in
daily social interaction. Constructions, in turn, may once again lead to interactions, in the sense
that constructions are ‘tested’ in interactions: do my ideas make sense, are they valid?
The construction of internal mental representations is one of the features of the process of
learning mathematics. We conceive the development of internal representations as a process of
signification (Kirshner & Whitson, 1997; Walkerdine, 1997). So we do not make a distinction
between an externally represented world and an internally representing world. Representation is
looked upon as a process in which new signs in a cyclic process of signification constantly
emerge. An internal representation (signifier) transforms and is the basis (signified) for the
construction of a new internal representation (signifier). Hence a person constructs internal
representations on the basis of internal representations.
The process of learning mathematics distinguishes itself from the process of learning other
school subjects to the extent that in mathematics, constructions – in the sense of internal
representations which children formulate based on knowledge gained through experience –
consistently show a closer correspondence with external representations than in other school
subjects (Cobb et al., 1987; Freudenthal, 1983). Children gain experience in the use of
measurements, numbers, ratios, and fractions and construct (intuitive) representations. In theory
these representations form a basis upon which the teacher can build, although this is not always
the case. Particularly this is not the case when children learn to operate with mathematical
symbols. Many errors are based on the default nature of natural language encoding processes, as
Kaput (1987) bas stated. Kaput discusses the well-known Student-Professors problem (Clement,
1982). At a certain university, for every 6 students there is one professor. Write an algebraic
equation that expresses the relation between the number of students and professors. Consistently
the natural language overrides the algebraic rules as is shown by the high error rates (40-80%)
across age and the predominance of the ‘6s = p’ error, typified by Kaput as the ‘reversal error’.
 The representations which the children construct concerning physical phenomena – also
known as preconceptions, misconceptions or intuitive ideas – generally deviate so far from actual
physical reality that they are useless as a basis for conceptualization. For example, children
associate energy with eating a Mars bar; in their minds, evaporation is the same as disappearing,
heat means feeling nice and warm, and light is a ray which goes from the eye to an object, rather
than the other way around (Van der Valk, 1989). These representations are useless in instruction,
but the teacher must be familiar with them in order to understand the problems that arise in
conceptualization. This applies as well to representation in other school subjects. For example,
when studying history, children have a great deal of trouble forming representations based on
their own experience. One child, for instance, regularly confused ‘Enlightenment’ with ‘more
light’.

Interaction

Realistic instruction in mathematics is not only constructive but also interactive. Several authors
have pointed out the importance of this interactive, or social, dimension of learning. Bishop
(1988) has argued to replace ‘impersonal learning’ and ‘text teaching’ with ‘mathematical
enculturation’, thereby emphasizing the relationship between education and culture. Pimm (1990)
uses the term ‘mathematical discourse’, while Salomon (1989) speaks of ‘cognitive partnership’.
Granott and Gardner (1994) constructed a theoretical framework of interaction, based on the
view of multiple intelligence approach. After their opinion the effect of interaction depends on
two dimensions. The first dimension is the relative expertise: non symmetric (‘parallel activity’
for instance) till asymmetric (‘apprenticeship’). The second dimension is the degree of
collaboration. ‘Scaffolding’ is for instance an example of collaboration of a high degree, while
‘imitation’ is in fact an independent activity (no collaboration).
Interactive teaching has also been called ‘cooperative learning’ (Slavin, 1986), ‘classroom
discourse’ (Cazden, 1988), ‘mutual instruction’ (Glaser, 1991) ‘guided construction of
knowledge’ (Mercer, 1995) and ‘interactive instruction’ (Treffers & Goffree, 1985). Bruner
(1986) a proponent of ‘discovery learning’ − learning on one’s own − in the 1960s, revised his
ideas several years ago: “My model of the child in those days was very much in the tradition of
the solo child mastering the world by presenting it to himself in his own terms. In the intervening
years I have come increasingly to recognize that most learning in most settings is a communal
activity, a sharing of the culture. It is this that leads me to emphasize not only discovery and
invention but the importance of negotiation and sharing.” (Bruner, 1986, p. 127). There is no
contradiction however between invention and sharing, the contrary is true: both activities
influence each other. If a person makes his own invention, it is worthwhile and even in many
cases necessary to discuss this invention. And this discussion is the basis for new inventions.
Nowadays, this view of (cognitive) development and learning is classified as social-
constructivism, a classification which meshes with the realistic approach to mathematics
instruction. In some studies (Driver, Asoko, Leach, Mortimer, & Scott, 1994; Roazzi & Bryant,
1994) is defended the point of view that learning and thinking always take place in a social
situation. Learning, they say, is situated learning (Kirshuer & Whitson, 1997), cognition is social
cognition. Bruner’s (1986) designation for the acquisition of knowledge is ‘negotiation of
meaning’. Not only words, concepts, gestures, and rituals, but also numbers, symbols, images,
visual and graphic representations, etc. have a whole range of meanings. In the case of children,
these meanings are frequently highly subjective. In response to the question “How old are you?”,
one child was heard to answer “I’m four, but when I ride in the bus I’m three” (in the
Nether1ands children under the age of four can ride public transportation free of charge).
Another child believes that when teachers roll the dice, they get double sixes more often than
children do. Teaching, says Bruner, means negotiating meaning. You say that zero is “nothing”,
but what then does zero degrees mean on the thermometer? A child does not believe that her face
has a surface. “Why not?” asks the teacher. “Because it isn’t length times width,” the child
responds. By applying the Socratic method, the teacher was gradually able to convince the child
that the ‘surface’ was not exclusively linked to the algorithm l × w. Bruner’s tribute to Vygotsky
(1977) is not at all surprising. According to the latter, a child’s higher psychic functions (such as
language and thought) first take shape as a social (interactive) activity and only later as an
individual activity. Language first functions as a means of communication; afterwards it
becomes internalized and serves an individual, self-regulatory function. One of the key concepts
in Vygotsky’s theory is that education should anticipate actual development. He refers in this
connection to the ‘zone of proximal development’, and it is this idea which inspired Bruner’s
‘negotiation of learning’. Both are concerned with interactive instruction, which Freudenthal
(1984) typifies as ‘anticipatory learning’ and Van Parreren (1988) as ‘developing education’.
 Realistic mathematics instruction is interactive, even though children must naturally be given
the chance to work independently. As demonstrated in studies carried out by Doise and Mugny
(1984), however, the point is that allowing children to experience various perspectives − in other
words, showing them that there are other children with other ideas about how to solve a
mathematics problem − will stimulate their thinking. Mechanistic (and individualistic)
mathematics instruction can exclude such experiences because children are required to comply
with the procedures given in their textbook. Discussion is restricted because the essence of
instruction lies in teaching irrefutable procedures. Realistic instruction, on the other hand, is
based on the exchange of ideas, not only, as in the past, between teacher and pupils, but also
between the pupils themselves. Interaction stimulates reasoning, using and analysing arguments,
thinking about own solutions and the solutions of others, so interaction reinforces the thinking
ability. Currently, social interaction in the classroom is receiving much attention where an
important line of research focuses on effects of small group work (Hiebert, 1992). It goes
without saying that there should be a ‘genuine’ occasion for discussion. That is why the point of
departure for realistic instruction is frequently a problem in context; again, this emphasizes how
tightly interwoven context and interaction are.
A simple example is the following: a teacher asks his class (6 and 7-year-olds) to think up as
many ways as they can of doing the sum 5 + 6. The children are allowed to discuss this among
themselves and together they came up with several methods: counting from 6 on up; adding 5 +
5 to reach 10 and then adding 1; counting the fingers on both hands and then adding 1; adding 6
+ 6 and subtracting1, etc. Teacher and pupils then discuss which method is the handiest and why.
This process leads the children to reflect spontaneously on their own actions: they are forced to
compare their methods with those of the other children and consider which is the best (this
example makes clear that in realistic mathematics instruction the students not only are
confronted with context problems, but with bare problem as well).

Reflection

According to Hiebert (1992), reflection or metacognition can be defined as the conscious

consideration of one’s experiences, often in the interests of establishing relationships between
ideas or actions. It involves thinking back on one’s experiences and taking the experiences as
objects of thought. With respect to terminology, reflection is seen most frequently in Russian
research reports (Davydov, Lompscher, & Markova, 1982; Nelissen & Tomic, 1996; Stepanov &
Semenov, 1985; Zak, 1984). The terms self-monitoring or self-regulation are also applied
(Glaser, 1991). It would be most tempting to spend a great deal of time discussing the many
questions, controversies and dilemmas which have arisen in the literature concerning the concept
of reflection − consider, for example, the discussion concerning the extent to which reflective
skills are general or contingent on context (Perkins & Salomon, 1989).
We generally do not reflect while performing a routine task, for the simple reason that there
is no cause to do so then. There is, however, reason for reflection whenever we are confronted
with a problem for which there is no immediate solution at hand. Reflection begins when we ask
ourselves how best to approach the problem: ‘Should I do it this way or that way?’ (planning).
Once we have set to work, other questions arise: “Is this working?” (self-monitoring), perhaps
even “Can I do it?” (self-evaluation). Other obvious questions are “Will this succeed?”
(anticipation) and, finally, “Am I happy with this?” (evaluation). If the solution turns out to be a
dead end, then we are forced to ask ourselves “Shouldn’t I try something else?” (consider
switching methods). These are, in brief, the most important elements of reflection during the
process of problem-solving. Reflection plays a significant role in learning to solve mathematical
problems, and indeed in human action in general. Through reflection students learn to analyse
their own actions critically and also become less dependent on their teacher. Their thinking
becomes more systematic, however this is not the case with all students. Some students must be
stimulated frequently. Reflection also allows them to investigate problem-solving methods and
procedures for general applicability, and increases the flexibility of their thinking. The most
important aspect, however, is that reflection builds self-confidence by allowing pupils to
discover what they really think and why they think it. Without this knowledge, every result
might seem – and in fact might very well be − serendipitous, an awareness that does little to
build up confidence: the pupil might not be so lucky the next time around.
That reflection is closely tied to the mathematical learning process and to mathematical
thinking can be deduced from the proposition, discussed above, that mathematization is a
constructive activity. This activity, in turn, is permanently linked to interaction, as we have seen.
We can imagine the connection between construction, interaction and reflection in the following
manner: constructive thinking implies that interaction takes place concerning our own
constructions (representations). We must naturally be able to test our own constructions and find
out how valuable they are. By exploring − and anticipating − the ideas and criticisms expressed
by others, we gain greater insight into our own ideas. Knowing what these ideas are and how we
ultimately came up with them is called reflection. We internalize the dialogue which we
originally conducted with others, turning it into a dialogue ‘with ourselves’. Reflection, thus, is
nothing less than ‘internalized dialogue’: from primarily inter-individual to intra-individual
activity. Through reflection, we continue to create new constructions, each time at a higher level.
In short, reflection is development.
There is a relationship between reflection and the process by which pupils solve
mathematical problems. In one study, the reflective thinking and mathematical problem-solving
skills of two groups of students were compared (Nelissen, 1987). One group had been taught
mathematics according to the realistic method (84 students) and the other according to the
mechanistic method (60 students). One striking result was that the students in the first group
were more flexible in their thinking than those in the second group, specifically because they
were better able to switch strategies whenever necessary. They were less likely to concentrate
purely on algorithmic solutions, and were able to develop strategies on the basis of their own
experience. They tended to check their own approach without prompting and were aware of their
own thought processes. In general, the children who were better able to solve problems were also
better at reflection, in the ‘realistic’ group this was 43% of the students, while in the control-
group this was 10% of the students.
A number of factors might serve to explain this close, positive relationship: (a) the school
curriculum followed by children in the experimental group was based on problem-solving within
a rich context. Instead of being given fixed, standard procedures to learn, pupils in these schools
were allowed to think up their own constructions. Through interaction they were encouraged to
reflect on their own approach. In this way, problem-solving and reflection were stimulated in
relation to each other. Note that the children were not given direct, separate training in reflection.
Research has shown that the training of functions in a separate programme has only a limited
effect (Derry & Murphy, 1986); (b) by commenting regularly on each other’s actions, the
‘realistic’ children were able to generate a reflective attitude which may have had a positive
impact on their problem-solving skills; (c) the children in the experimental group were taught the
concepts, models and procedures they needed to solve problems and engage in reflection. To be
able to reflect on a specific subject, they needed to acquire domain-specific knowledge; (d)
reflection will only prove beneficial after children have come to view the actions they are
reflecting on as meaningful. The children in the ‘realistic’ group found mathematics and
problem-solving a meaningful activity. They were therefore more inclined to reflect on problem-
solving than the children in the ‘mechanistic’ group. Children in the latter group saw little reason
to apply their own reality to learning mathematics, because this reality was continually
supplanted by prescribed standard algorithms. A study conducted by Stepanov and Semenov
(1985) revealed that in order to be able to reflect on the process of problem solving, children
must first see their own actions during this same process as meaningful. Meaning must therefore
be given due attention during instruction, if children are to find reflection a meaningful activity.
Reflection, in turn, is vital if the mathematization process is to run smoothly.

Solving mathematical problems

One important objective in mathematics instruction is that children be able to apply the concepts
and skills they have acquired with reasonable success. Problem-solving is considered by many to
be one of the most important areas of application. That is why so many researchers are interested
in whether, and if so, how children solve mathematics problems. Where upon are processes of
problem solving based, on declarative knowledge or on knowledge of procedures?
The respective opponents and proponents of ‘declarative representations’ and ‘procedural
representations’ have been engaged in a vehement debate since the early 1970s (Gardner, 1987).
Adherents of the first believe that the knowledge base is the most important factor in problem-
solving, while adherents of the second relate success largely to the use of procedures and
strategies. In the 1980s the dispute concerned whether such knowledge or such (reflective)
procedures were general or domain-specific. This controversy led to yet another split in both
camps. Although the debate rages on, its resolution seems to be in sight, specifically because
human thought is increasingly being characterized as modular. Learning is therefore by no means
a ‘content blind’ process; neither, according to Gardner (1987) are there such things as ‘general
cognitive architectures’, as Piaget (1976) suggested. Several leading authors appear to share this
opinion: Bonner (1990), Resnick and Klopfer (1989), and Schoenfeld (1989). All of these above
mentioned researchers tend, albeit from different backgrounds, to maintain that problem-solving
(in mathematics) will be most successful when based on a well-organized selection of domain-
specific knowledge, but that the use of procedures which are (once again) domain-specific is also
indispensable. Experts tend to use domain specific procedures and principles while novices are
more likely to choose general strategies. For this reason novices often fail to solve the problems
(Caillot, 1991).
Problem-solving in mathematics, then, is also characterized by a specific mathematical
approach, involving the use of domain specific concepts, tools (procedures) and ways of thinking.
A child who has not mastered this approach will have difficulty when solving mathematical
problems. A few examples follow. A classroom of six- and seven-year-olds were given the
following problem: 2 friends live next door to each other, one at number 3 and one at number 5.
How many houses do they live in? The children answered: 8 houses. Numbers mean little to
these children, except that they can be added up. Assigning a meaning to numbers is an
important component of a mathematical approach. Many children (eleven or twelve years old)
have difficulty solving the following type of problem: a walkman costs $150 after the price has
gone up by 20%. What was the original price? The answer most frequently given is: 20% of
$150 is $30; the original price was therefore $120. This answer is, of course, incorrect. Insight
into this problem can, however, be provided by means of a diagram showing that $150 does not
equal 100% but 120%:

15 15 15 15 15 to be added: 15

20% 20% 20% 20% 20% 20%

5 5 is 100%; 6 5 then is 120%. So 6 5 × $150 = $125.

Schematization − or visualization − is an important mathematical strategy, a tool for solving

problems. Other such tools are: estimating, simplifying problems, testing, changing perspectives
and conducting a thought experiment. Gravemeijer (1988) has classified mathematization tools
according to characteristics derived from mathematics itself. For example, structuring and
generalizing are related to the category ‘generality’. Proofs and predictions are connected to
‘certainty’ because this is for a mathematician very important, symbolization and formalization
belong to ‘precision’ and reduction and constructing algorithms to ‘conciseness’.
 Resnick and Klopfer (1989) identify knowledge that plays an important role in problem-
solving as ‘organizing schemata’, concepts which are ‘powerful’ and must be actively acquired.
In other words, it is rich, flexible, ‘generative knowledge’. This knowledge, or specialist
knowledge, forms the basis upon which we can construct the first representation of a problem
which we must solve. This representation is the starting point for a successful problem-solving
process. Bransford, Sherwood, Vye, and Rieser (1989) warn against the danger of rote
knowledge or, as they put it, ‘inert knowledge’. Because children do not consider such formal
knowledge ‘real’ or meaningful, they will not be able to apply it or only do so blindly,
particularly in mathematics. Children acquire inert knowledge in mathematics when they are
forced to learn formulas such as ‘To divide by a fraction, multiply by its opposite’ or to do plain
problems involving meaningless numbers (leading to the type of problem discussed in the
example above). In this connection, realistic mathematics instruction makes use of the models,
schemata and concepts − called ‘conceptual models’ by Lesh (1985) − which form the core of
the mathematical approach.
This approach is similar to Davydov’s (1977) idea of teaching children to work with
theoretical concepts − which can be seen as concepts essential to a specific discipline − instead
of concepts based on observation or empiricism. In this connection, Davydov has argued for the
formation of theoretical thinking, similar in certain respects to Freudenthal’s (1984) development
of mathematical thinking or attitude. In fact, the two are so similar that both have pointed out the
same flaws in the empirical, inductive teaching approach. In this approach, knowledge comes
into being through observation or empiricism. Here we can recognize the influence of the
empiricist school, which, as we have seen, places very little emphasis on vertical
mathematization − which is precisely what Freudenthal and Davydov do wish to emphasize.
Mathematical or theoretical insights are required in order to be able to understand reality
correctly (including learning tools such as the number line or abacus). Here, however, the
similarity between Freudenthal and Davydov ends. According to Freudenthal, Davydov
introduces theoretical concepts too early on in education; moreover, Freudenthal rejects the
distinction made by Davydov between theoretical and empirical concepts. In Freudenthal’s view,
theory is always inherent in empiricism; all action is implicitly theory bound.
Research into solving mathematics problems often focuses on initial mathematics problems
and word problems. See for example De Corte, Verschaffel, and Greer (1996), Span, De Corte,
and Van Hout-Wolters (1989) and Verschaffel and De Corte (1990) for a report on studies
carried out on learning and problem-solving in mathematics. Following in the footsteps of Riley,
Greeno and Heller (1983), Verschaffel and De Corte (1990) explored which internal
representations children formed as a result of problems which they were given to solve, and
which role the semantic features of these problems played in this. They also wished to discover
how these representations formed the basis for actions, in particular problem-solving procedures.
Their research revealed that semantic factors played a role in forming problem representations
Many children did indeed tend to take their lead from the meaning of words (some of which
were printed by chance) in the text (‘together’ or ‘with each other’) and to base their solution
procedures on these words. It was also shown that many verbal, nonpropositional, grammatical
characteristics influenced the choice of procedure and not only cognitive schemata, as authors
such as Resnick (1983) and Riley et al (1983) have suggested. Finding the correct solution, these
researchers believe, depends on the formation of adequate representations constructed from part-
whole schemata.
There are, however, several objections which might be raised (Van Luit, 1994) to this
emphasis upon the part-whole schema. To begin, there is no relationship with a child’s previous
experiences, such as counting (Van Mulken, 1992). Second, this schema is based exclusively on
the cardinal interpretation of numbers, whereas in truth numbers may appear in other forms: 9 is
6 + 3, but 9 is also 3 × 3 or the root of 81. Third, a child gets into trouble if he or she comes
across a problem which does not contain a part-whole relationship; for example, John is 5 years
old and his friend is 6.
A more general comment on research into solving word problems is that the influence
exerted by semantic structural features has been given too much emphasis, at the expense of
studying the influence which the nature and the size of numbers have on the choice of solution
procedure (Van Mulken, 1992; Verschaffel & De Corte, 1990). This is known as ‘number
sensitivity’ An example is: 62 - 18 = ? If we take account of the nature of the numbers in this
problem, the obvious strategy is to subtract 20 first and then add 2 The problem 62 - 33 =?,
however, requires a different approach, for example: 63 - 33 = 30, 30 - 1 = 29. For a lot of
children this is not an easy problem, because it supposes much flexibility in thinking. In teaching
children to solve word problems, one should emphasize neither mastery of procedures (such as
the use of the part-whole schema) nor semantic features, but rather the structure of the problem,
specifically the structure of the numbers. Proponents of realistic mathematics instruction argue
for flexibility in choosing problem-solving methods or learning to choose them.
Some of the above mentioned researchers maintain that training in meta-cognitive skills –
such as planning the course of a solution, setting up schemata, guessing the solution in advance –
can simplify the approach taken to problems. However, in none of the studies was training
conclusively shown to be successful, although progress was noted. Some researchers point out,
no doubt correctly, that the weaker children lacked a certain knowledge background It was
remarkable that the children were able to master a heuristic (for example, making an estimate),
but that this did not automatically lead to their choosing the required, formal operation.
In problem solving the context can be important. The function of contexts is to elicit
knowledge which children have gained through experience and which they can use once again in
forming internal problem representations. This is not, however, a hard and fast rule. Sometimes
meanings which have been acquired through experience are directly contrary to the mathematical
meaning of concepts (and, as we remarked earlier, in physics this is the rule rather than the
exception (Keil, 1989). Walkerdine (1988) drew attention to this in her series of carefully
conducted observations. An association with the child’s experience, she claims, means an
association with contexts, and these are highly domain-specific. We cannot simply assume,
therefore, that ‘transfer’ is affected simply by the insertion of mathematical relations into a
‘meaningful context’. The author illustrates this by giving the following example. Children first
learn the concept ‘more’ in a pedagogical situation: ‘Just two mouthfuls more, mmm!’ In
mathematics, however, ‘more’ is the opposite of ‘less’, and that is an entirely different concept.
There are, then, two entirely different ‘discursive practices’; in other words, it is not always the
case that children solve problems better by using knowledge based on their own experience.

Discussion and conclusions

There have been radical changes in the approach to mathematics instruction in the past few years.
These changes came about because mathematicians began to view their own discipline
differently, leading to new research on teaching methodology. This research was supported by
new developments in educational psychology. Glaser (1991) analyses these developments. He
points out the tendency to relate learning and thinking to specific domains. This approach has
been defended by various authors. For example, in a series of studies (Keil, 1989) it is shown
that indeed it is plausible that concepts develop largely in specific domains. Research has also
demonstrated the importance of reflection.
Experts − adults or children − develop skills in order to plan and monitor their actions and
predict what the results of their efforts will be. In an experiment, Glancey (1986) introduced the
knowledge representation (including heuristic rules) of experts to a group of pupils. On the basis
of these representations, the pupils learned to formulate hypotheses, recognize errors and, in
particular, to better organize their knowledge base. Glancey’s most original idea was to have the
pupils observe experts and question them concerning their methods. The observation strategies
and the questions were analyzed in advance, and an analysis was also done of how knowledge
might be restructured during the learning process. Here interaction and reflection go hand in
hand. The idea of learning as a process of continual restructuring, of increasing architectonics
and deeper insight, has also been expounded by White and Frederiksen (1986). They furthermore
emphasize establishing a link with the pupils’ naive, intuitive models, as in mathematics
instruction.
Constructivism is another tendency which Glaser (1991) has frequently come across in
analyzing research reports. Although the emphasis in the past was on monitoring the pupils’
learning processes, nowadays the pupils has more control over his own learning environment,
and many studies have attempted to gain insight into how pupils construct their learning
environment in order to be able to learn something, A new image of the pupil is coming into
being; they are no longer ‘good boys and girls’ who learn everything by rote, but children
motivated to explore and seek explanations.
It was not only among Anglo-Saxon cognitive psychology that is undergoing significant
changes; the Russian Cultural Historical School has long been concentrating on research themes
related to the new developments in mathematics instruction. The main focus is on reflective
thinking (Stepanov & Semenov, 1985; Zak, 1984), interaction (Davydov et al., 1982), and the
development of domain-specific concepts (Davydov, 1977), while education and instruction are
in essence seen as the active interrelation of symbolic systems and meanings which a culture has
brought forth (Leont’ev, 1980; Van Oers, 1987). For a more thorough comparison between the
concepts presented by the Russian Cultural Historic School and realistic teaching methodology,
readers are referred to Nelissen and Tomic (1995, 1996).
 So, in the past few years a number of interesting new themes relevant to mathematics
instruction have received a great deal of attention. We have argued that learning is a process
which rests upon children’s own constructive activity. Learning takes place in a social context
(Bruner, 1996; Slavin, 1986). Mathematical ideas are not merely abstract; they are contained
within language and concepts. Learning is a process in which the child masters its cultural
heritage, by learning particular sets of symbols. If children are able to put their ideas into words,
they will have a better grasp of their own way of thinking. We have also stated that when
children are able to put their ideas into words, their teacher will gain greater insight into their
thought processes. If their constructions prove to be unusable, the teacher can confront the
children with alternative approaches, for example through problem-oriented questioning. When
children discuss their ideas with one another (Mercer, 1995), they not only have to state their
opinions more concisely, they also have to listen, think along with others and try to understand
what the other children actually mean to say. Learning − and cognitive development − is
increasingly viewed as a process in which metacognition (or reflection) fulfills a regulatory
function.
Solving mathematics problems requires learning domain-specific rather than general
knowledge. This knowledge is well-structured and flexible, and encompasses a knowledge of
both content and procedures and reflective knowledge. In mathematics, tools and modes of
thinking that typify mathematics should be maintained, and used to solve problems. The
overriding concern is to maintain a flexible choice of tools, and that choice requires domain-
specific reflection. To be able to reflect, however, knowledge of content is once again necessary;
one can only reflect on the use of tools, strategies and concepts if one knows them. By now many
of these new ideas have filtered through to the practice of teaching mathematics, in part because
new programmes are being developed, implemented and supervised which are based on the
realistic approach, and in part because teacher training now focuses on the new realistic didactic.
It is, ultimately, the teachers themselves who must put these new ideas into practice.
Although there is a high degree of consensus among researchers in mathematics instruction,
particularly on initial mathematics problems and word problems, up to the present theory has
preceded carefully collected empirical data. If we also agree, on the basis of research findings,
that construction, interaction and reflection are essential for learning mathematics, then the
practice of mathematics instruction should be altered radically. Teachers are being asked to
master a new approach to instruction and to their pupils. Among other things, this means a new
approach to testing, to explaining, to cooperating and discussing, to working independently, to
thinking intuitively, to understanding and developing concepts.

References

Bishop, A.J (1988). Mathematical enculturation A cultural perspective on mathematics

education. Dordrecht: Kluwer.
Bonner, J. (1990). De implicaties van de cognitieve theorie voor het ontwerpen van instructie:
Een heroverweging [Implications of cognitive theory for instructional design: A
reconsideration]. In J.S. ten Brinke (Ed.), Paspoort onderwijspraktijk (Vol. 3, pp. 453-463).
Houten: Bohn Stafleu Van Loghum.
Bransford, J., Sherwood, R., Vye, N., & Rieser, J (1989). Leren denken en probleem oplossen
[Learning to think and problem solving]. In G. Ackers (Ed.), Paspoort Onderwijspraktijk (pp.
155.190). Houten: BohnStafleu Van Loghum.
Bruner, J. (1986). Actual minds, possible worlds. Cambridge, MA: Harvard University Press.
Bruner, J. (1996). The culture of education. Cambridge: Harvard University Press.
Caillot, M. (1991). Learning thinking through the new technologies. In S. Maclure & P. Davies
(Eds.), Learning to thin: Thinking to learn. Oxford: Pergamon.
Cazden, C. (1988). Classroom discourse The language of teaching and learning. New York:
Heineman.
Clement, J. (1982). Algebra word problems solutions: Thought processes underlying a common
misconception. Journal for Research in Mathematics Education, 16, 16-30.
Cobb, P. (1994). Where is the mind? Constructivist and sociocultural perspectives on
mathematical development. Educational Researcher, 23, 7, 13-20.
Cobb, P., Yackel, E., & Wood, T. (1992). A constructivist alternative to the representational
view of mind in mathematics education. Journal for Research in Mathematics Education, 14,
83-95.
Cobb, P., Gravemeijer K., Yackel E., McClain K., & Whitenack J. (1997). Mathematizing and
symbolizing: The emergence of chains of signification in one first-grade classroom. In D.
Kirshner & J.A. Whitson (Eds.), Situated cognition (pp. 151-233). Hillsdale, NJ: Lawrence
Erlbaum Associates.
Conftey, J. (1981). Conceptual change analysis: Implications for mathematics and curriculum.
Toronto: Wiley.
Confrey, J. (1985). Towards a framework for constructivists instruction. In L. Streefland (Ed.),
Proceedings of the Ninth Conference for the Psychology of Mathematics Education (Vol. I,
pp. 477-483). Noordwijkerhout: PME.
Davydov, V.V. (1977). De introductie van het begrip grootheid in de eerste klas van de
basisschool [Introduction of the quantity concept in first grade elementary school]. In CF.
van Parreren & J.M.C. Nelissen (Eds.), Rekenen Teksten en analyses Sovjetpsychologie (Vol.
2, pp. 1-61). Groningen: Wolters- Noordhoff.
Davydov, V.V., Lompscher, J., & Markova, A.K. (1982). Ausbildung der Lerntätigkeit bei
Schülern [Development of learning in pupils]. Berlin: Volk und Wissen.
De Corte, E., Greer B., & Verschaffel L. (1996). Mathematics teaching and Learning. In D.
Berliner & R. Calfee (Eds.), Handbook of educational psychology (pp. 491-549). New York:
MacMillan.
Derry, S.J., & Murphy, D.A. (1986). Designing systems that train learning ability: From theory
to practice. Review of Educational Research, 56, 1, 1-40.
Doise, W., & Mugny, G. (1984). The social development of the intellect. Oxford: Pergamon.
Driver, R., Asoko, H., Leach, J., Mortimer, E., & Scott, P. (1994). Constructing scientific
knowledge in the classroom. Educational Researcher, 23, 7, 5-13.
Freudenthal, H. (1983). Didactical phenomenology of mathematical structures. Dordrecht:
Reidel.
Freudenthal, H. (1984). Appels en peren / wiskunde en psychologie Gebundelde opstellen
[Apples and pears / mathematics and psychology: Collected essays]. Apeldoorn: Van
Walraven.
Gardner, H. (1987). The mind’s new science A history of the cognitive revolution. New York:
Basic.
Glancey, W.J. (1986). From Guidon to Neomycin and Hercules in twenty short lessons: ONR
final report 1979-1985. AI Magazine, 7, 40-60.
Glaser, R. (1991). Wedergeboorte van de leertheorie in onderwijskundig onderzoek [Rebirth of
learning theory in educational research]. In J.S. ten Brinke (Ed.), Paspoort onderwijspraktijk
(pp. 1.1-1.32). Houten: Bohn Stafleu Van Loghum.
Goffree, F., Freudenthal, H, & Schoemaker, G. (1981). Wiskundeonderwijs. Een micro-
didactische beschouwing over het thema ‘begrijpen’ [Mathematics education. A micro
pedagogical consideration on ‘understanding’]. In H.P .M. Cremers (Ed.), Losbladig
onderwijskundig lexicon (pp. 1200.3-1200.77). Alphen a/d Rijn: Samsom H.D. Tjeenk
Willink.
Granott, N., & Gardner, H. (1994). When minds meet: interaction, coincidence and development
in domains of activity. In R.J. Sternberg & R.K. Wagner (Eds.), Mind in contexts.
Interactionist perspectives on human intelligence (pp. 133- 152). Cambridge, MA:
Cambridge University Press.
Gravemeijer, K.P.E. (1988). De grondslagen van het programma Rekenen en Wiskunde
[Foundations of the programme Arithmetic and Mathematics]. Rotterdam: OSM.
Gravemeijer, K. (1997). Commentary solving word problems: A case of modelling? Learning
and Instruction, 7, 389-307.
Greer, B. (1997) Modelling reality in mathematics classrooms: The case of word problems.
Learning and Instruction, 7, 293-307.
Hiebert, J. (1992). Reflection and communication: Cognitive considerations in school
mathematics reform. International Journal of Educational Research, 17, 439-456.
Kaput, J.J. (1987). Towards a theory of symbol use in mathematics. In: C. Janvier (Ed.),
Problems of representation in the teaching and learning of mathematics (pp. 159-197).
Hillsdale, NJ: Lawrence Erlbaum.
Keil, F.C. (1989). Concepts, kinds and cognitive development. Cambridge, MA: MIT.
Kirshner, D., & Whitson J.A (Eds.) (1997). Situated cognition. Hillsdale, NJ Lawrence Erlbaum.
Lakatos, I. (1976) Proofs and refutations. The logic and mathematical discovery. London
Cambridge University Press.
Leont’ev, A.N. (1980). Activiteit als psychologisch probleem [Activity as a psychological
problem]. Pedagogische Studiën, 57, 324-343.
Lesh, R. (1985). Conceptual analysis of mathematical ideas and problem solving processes. In L.
Streefland (Ed.), Proceedings of the Ninth Conference for the Psychology of Mathematics
Education (Vol. 2, pp. 73-97). Noordwijkerhout: PME
Lesh, R., Behr M., & Post T. (1987). Rational number relations and proportions. In C. Janvier
(Ed.), Problems of representation in the teaching and learning of mathematics. Hillsdale, NJ:
Lawrence Erlbaum.
Lo, J.J., Grayson, H., Wheatly, J., & Smith, A. (1990). Potential mathematics learning
opportunities in grade three class discussion. In G. Booker, P. Cobb, & T. De Mendicuti
(Eds.), Proceedings of the Fourteenth Conference for the Psychology of Mathematics
Education (pp. 113-121). Mexico: PME.
Mercer, N. (1995). The guided construction of knowledge. Clevedon, Phil: Multilingual Matters.
Nelissen, J.M.C. (1987). Kinderen leren wiskunde. Een studie over constructie en reflectie in het
basisonderwijs [Children learn mathematics. A study concerning construction and reflection
in elementary school education]. Gorinchem: De Ruiter.
Nelissen, J.M.C., & Tomic, W. (1995). Reflection in education with an example from
mathematics. Educational Practice and Theory, 17, 53-70.
Nelissen, J.M.C., & Tomic, W. (1996). Reflections in Russian educational psychology.
Educational Foundations, 10, 35-57.
Papert, S. (1980). Mind storms. Children, computers and powerful ideas. New York: Basic.
Piaget, J. (1976). Genetische epistemologie [Genetic epistemology]. Meppel: Boom.
Perkins, D.N., & Salomon, G. (1989) Are cognitive skills context bound? Educational
Researcher, 18, 1, 16-25.
Pimm, D. (1990). Certain metonymic aspects of mathematical discourse. In G. Booker, P. Cobb,
& T. de Mendicuti (Eds.), Proceedings of the Fourteenth Conference for the Psychology of
Mathematics Education (pp. 129-136) Mexico: PME.
Presseisen, B.Z., Sternberg, R.J., Fischer, K. W., Knight, C.C., & Feuerstein. R (1990). Learning
and thinking styles: Classroom interaction. Washington: NEA.
Resnick, L.B. (1983). A developmental theory of number understanding. In H.P. Ginsburg (Ed.),
The development of mathematical thinking (pp. 109-151). New York: Academic Press.
Resnick, L.B., & Klopfer, L.E. (1989). Toward the thinking curriculum: Current cognitive
research. Alexandria, VA: ASCD.
Reusser, K., & Stebler R. (1997). Every word problem has a solution. The social rationality of
mathematical modeling in schools. Learning and Instruction, 7, 309-327.
Riley, M.S., Greeno, J.G., & Heller, J.I. (1983). Development of children’s problem-solving
ability in arithmetic. In H.P. Ginsburg (Ed.), The development of mathematical thinking (pp.
153-196). New York: Academic Press.
Roazzi, A., & Bryant, P. (1994). Social class, context and cognitive development. In P. Light &
G. Butterworth (Eds.), Context and cognition. New York: Harvester Wheatsheaf.
Salomon, G. (1989). Technology for the mind: Partner and nourisher. In P. Span, E. De Corte, &
B. van Hout-Wolters (Eds.), Onderwijsleerprocessen: Strategieën voor de verwerking van
informatie (pp 67-84). Amsterdam: Swets & Zeitlinger.
Saxe, G.B. (1988). Candy selling and math learning. Educational Researcher, 17, 1,14-21.
Schoenfeld, A.H. (1989). Teaching mathematical thinking and problem solving. In L.B. Resnick
& L.E. Klopfer (Eds.), Toward the thinking curriculum: Current cognitive research (pp. 83-
104). Alexandria, V A: ASCD.
Slavin, R.E. (1986). Developmental and motivational perspectives on cooperative learning. San
Francisco, CA: AERA.
Slavin, R.E. (1996). Education for all. Amsterdam: Swets & Zeitlinger.
Span, P., De Corte, E., & Van Hout-Wolters, B. (1989). Onderwijsleerprocessen. Strategieën
voor de verwerking van informatie [Teaching learning processes. Strategies for information
processing]. Amsterdam: Swets & Zeitlinger.
Steffe, L., Cobb, P., & Von Glazersfield, E. (1988). Construction of arithmetical meanings and
strategies. New York: Springer Verlag.
Stepanov, S.J., & Semenov, I.N. (1985). Problemy psichologičeskogo izučenija refleksii I
trorčestva [Psychology of reflection: problems and investigations]. Voprosy Psichologii, 3,
31-40.
Streefland, L. (1991). Fractions in realistic mathematics education. Dordrecht: Kluwer.
Treffers, A. (1987). Three dimensions. A model of goal and theory description in primary
mathematics education Wiskobas. Dordrecht: Reidel.
Treffers, A. (1991). Didactical background of a mathematics programme for primary education.
In L. Streefland (Ed.), Realistic mathematics education in the primary school (pp. 21-57).
Utrecht: Freudenthal Institute.
Treffers, A., & Goffree, F. (1985). Rational analysis of realistic mathematics education. In L.
Streefland (Ed.), Proceedings of the Ninth Conference for the Psychology of Mathematics
Education (Vol. 2, pp. 97-123). Noordwijkerhout: PME.
Van den Heuvel-Panhuizen, M. (1996). Assessment and realistic mathematics education. Utrecht:
Freudenthal Institute.
Van der Valk, T. (1989). Het energie-begrip van leefwereld naar natuurkunde [The concept of
energy from subjective view to physics]. Didaktief, 19, 4, 22-25.
Van Luit, J.EH. (Ed.) (1994). Research on learning and instruction of mathematics in
kindergarten and primary school. Doetinchem: Graviant.
Van Mulken, F. (1992). Hoofdrekenen en strategisch handelen [Mental arithmetic and strategic
action]. Leiden: University Press.
Van Oers, B. (1987). Activiteit en begrip. Proeve van een handelings-psychologische didactiek
[Activity and understanding. Exemplar of an action psychological pedagogy]. Amsterdam:
Vrije Universiteit.
Van Parreren, C.F. (1988). Ontwikkelend Onderwijs [Developing Education]. Leuven: Acco.
Verschaffel, L., & De Corte, E. (1990). De techniek van de computersimulatie in het onderzoek
naar het leren oplossen van vraagstukken over optellen en aftrekken [The method of
computer simulation in research into learning to solve addition and subtraction problems].
Pedagogisch Tijdschrift, 15, 259-269.
Verschaffel, L., De Corte E., & Borghart I. (1997). Pre-service teacher’s conceptions and beliefs
about the role of real world knowledge in mathematical modelling of school word problems.
Teaching and Instruction, 7, 339-359.
Von Glazersfeld, E. (1991). Radical constructivism in mathematics education. Dordrecht:
Kluwer.
Vygotsky, L.S. (1977). Denken und Sprechen [Thought and language]. Frankfurt am Main:
Fischer.
Walkerdine, V. (1988). The mastery of reason. London: Routledge.
Walkerdine, V. (1997). Redefining the subject in situated cognition theory. In Kirshner, D. & I.A.
Whitson (Eds.), Situated cognition. Hillsdale, NJ: Lawrence Erlbaum.
White, B.Y., & Frederiksen I.R. (1986). Progressions of quantitative models as a foundotion for
intelligent learning environments. Cambridge, MA: Bolt, Beranek & Newman.
Whitney, H. (1985). Taking responsibility in school mathematics education. In L. Streefland
(Ed.), Proceedings of the Ninth Conference for the Psychology of Mathematics Education (pp.
123-142). Noordwijkerhout: PME.
Zak, A.Z. (1984). Razvitije teoretičeskogo myš1enija mladšich školnikov [Development in
theoretical thinking in young children]. Moskou: Pedagogika.