School Didactics And Learning: a school Didactic Model Framing An Analysis of Pedagogical Implication of Learning Theory

Elements of a theory of problem solving
(1958):
The heart of [our] approach is describing the behavior of the system by a well specified program,
defined in terms of elementary information processes … Once the program has been specified, we
proceed exactly as we do with traditional mathemathical systems. We attempt to deduce general
properties of the system from the program (the equations); we compare the behavior predicted from
the program (from the equations) with actual behavior observed…[and] we modify the program when
modification is required to fit the facts.
This view was developed in a later publication (Newell & Simon, 1972). The aim has been to develop
computer simulations of such steps as human beings take in the course of solving problems. The methods to
achieve this aim are of course different within the two traditions. It is this common modelling interest of
cognitive psychology and AI which has also been called cognitive science.
This way of carrying out empirical research within cognitive science has, according to Winograd (1980,
p. 226), followed the pattern described below. What is simulated is human cognition:
1. The scientist observes some recurrent pattern of interaction of an organism;
2. He or she devises some formal representation (for example, a set of generative rules or a “schema”)
that characterizes the regularities;
3. The organism is assumed to “have” the representation, in order to be able to exhibit the regularities;
4. (Depending on the particular subfield). The scientist looks for experiments that will demonstrate the
presence of the representation, or designs a computer program using it to see whether the behaviour can
be generated by the program.
A final point in specifying the computationalist approach or cognitive science is the de-emphasizing of the
influence of emotions, cultural and historical factors (Gardner, 1987, p. 6). Further, the background
or present context in which particular cognitive processes take place is not taken into account when these
processes are explained. I have pointed out that it is argued that de-emphasizing the context, historical or
social, present or past, is only a methodological issue. Accordingly, representatives of the approach claim
that they do not deny the possible importance or influence which these surrounding factors may have on
cognitive activities, but they are rather regarded as undeniably disturbing elements. It is concluded that “…
an inclusion of them at this point would complicate the cognitive science endeavour” (Gardner, 1987, p. 6).
What appears here is a view that it is possible to study human cognition as such without regarding the
context in which this activity occurs.
Theories on Different Levels
A final way of structuring the field is to focus on the variation of how different subparts within cognitive
psychology define their field.
Sometimes a division between the following levels is suggested:
104
SCHOOL DIDACTICS AND LEARNING

---

1. Intentional psychology focuses upon how an individual deals with his or her environment
(phenomenology, representational theories);
2. Sign-level psychology seeks to develop internal processing models of cognition (computationalism,
functionalism);
3. Physical-level psychology tries to explain mental events in purely physicalist or neuroscientific terms
(eliminative materialism, typeidentity theory, physical theories, connectionism).
It has been suggested that in certain respects the different psychological approaches do not compete.
Pylyshyn (1984) again argues that it is possible to identify three levels of psychological theory within
cognitive science which all represent a computational view of mind. They are:
(a) Representational theories;
(b) Information-processing theories; and
(c) Physical theories.
Representational theories express how the external world is captured by mental models.
2
Information-
processing theories deal with the human mental language (mentalese, cf. Fodor, 1975), trying to define it.
Physical theories would explain this inner language in terms of brain processes. Fodor (1975) has argued
that information-processing theories must intervene between physical and representational theories because
the representation must be handled by some general language. 
The differences between the ways of structuring the field are thus not overwhelming and it seems that the
cognitive science approach supports these structures with minor variations (Bechtel, 1988, pp. 54–78).
Cognitivism and the Level of Analysis in this Study
The two problems discussed, i.e. the ontological and the epistemological, have been emphasized in various
ways within cognitive science. Within representationalist theory the epistemological problem is easily
identified (i.e. how the relation between the content of mind and the external world is specified). But the
ontological problem is also discussed clearly; a representational level of description is defended against e.g.
eliminative materialism. Within computational and connectionist theory again, the ontological problem, i.e.
how we should describe the content of awareness, is primarily dealt with.
It is hoped that the previous discussion has shown how difficult it has been to construct an object of
analysis in this study that could be kept clearly distinct from other approaches. Although there are
differences among different directions of the cognitivist school of thought, it is considered possible to deal
with these directions as representing one and the same framework. The main interest concerns how the
cognitive paradigm in general and the cognitivist information-processing approach in particular, as well as
what is called representational theories, explain conceptual learning. Thus, the study does not include
computational learning theory or connectionist theories (see Anthony & Biggs, 1992; Kivinen, 1992).
Instead, attention is paid to approaches of cognition which view learning as conceptual change and share the
view of the individual as receiving and manipulating information.
COGNITIVISM AND THE THEORY OF LEARNING
The cognitive theory of learning in general has for the last three decades focused on acquisition of
knowledge structures rather than behaviour. Behaviour has been seen as something resulting from what a
5. OBJECT OF ANALYSIS
105

---

subject has learned, i.e. behaviour as such is not learned. During the same period attention has shifted from
investigating changes in the environment which would provide for appropriate responses to focusing on
changes within the learner as well as on the learner’s activity. Nonetheless cognitive psychology makes use
of behavioural data, but these data are used to deduce mental processes or structures that are thought to
result in behaviour.
It is not very easy to describe how learning is conceived of within cognitive psychological research. This
is partly due to the problem of how the field should be limited, as we noted in the previous chapter. Another
reason is that there does not seem to exist any elaborate and generally accepted theory of learning within
cognitive psychology and cognitive science. This is the conclusion several leading researchers in the field
e.g. Fodor (1980, p. 149), Bereiter (1985, p. 201) and Ohlsson (Sandberg & Barnard, 1991, p. 139) have
arrived at:
There literally isn’t such a thing as the notion of learning a conceptual system richer than the one that
one already has; we simply have no idea of what it would be like to get from a conceptually
impoverished to a conceptually richer system by anything like a process of learning.
It seems to be generally agreed that there is no adequate theory of learning—that is, no adequate
theory to explain how new organizations of concepts and how new more complex cognitive
procedures are acquired.
There is a widespread belief that…we many theories of learning. If you put some reasonable
constraints on what you are willing to call a theory of learning that is not true at all.
One reason for this situation is that cognitivist psychology has developed models describing performance of
tasks rather than learning (Glaser & Bassok, 1989, p. 634):
Over the past quarter of a century, cognitive research has focused primarily on the analysis of
competence. Studies of memory, language, and problem solving have examined the nature of
performance and the outcomes of learning and development… The least developed component of
instructional theory is explication of the process of learning—a contrast indeed to behavioral
psychology, where learning was of major concern…[T]he study of the transition processes that a
theory of learning must account for has been a depressed endeavor until recent years.
Simon (1979, p. 981) has defended the situation by claiming that it is natural to first explain what it is to
have knowledge before we try to explain how this knowledge is reached. Norman’s (1987, pp. 328–329)
position is similar:
Today the study of learning is not considered a central part of either psychology or artificial
intelligence. Why? Perhaps the understanding of learning requires knowing about problems or
representation, of input (perception), of output (performance), and of thought and inference.
In principle this is the same argument that was presented by Alain Newell and Herbert Simon in 1972; it is
difficult to understand what learning is about without having an idea of the knowledge states between which
a transition is supposed to occur. The performance system must be known.
Yet Norman (1987) admits that one reason why “so little is known about learning” is in part due to “lack
of trying”. There is an ongoing change in interest in favour of learning theory in contemporary cognitivism
(see e.g. Glaser, 1990).
106
SCHOOL DIDACTICS AND LEARNING

---

However it is interesting to note that even though learning has not been of prime interest in cognitive
psychology for the last few decades, many researchers are very explicit on what an adequate theory of
learning 
should
explain. To them, the main goal of learning theory seems to be to explain what it means for
an individual to “get from a conceptually impoverished situation to a conceptually richer system” (Fodor,
1980) and “how new organizations of concepts and how new more complex cognitive procedures are
acquired” (Bereiter, 1985). What learning theory thus should explain is the changes of the conceptual system
and cognitive procedures. Stellan Ohlsson (in Sandberg & Barnard, 1991, p. 139) also has a clear view of
what a learning theory should be about:
To me a learning theory should be saying: here are the mechanisms that are responsible for any kind of
cognitive change, here are the conditions under which these are triggered. Here are the knowledge
structures that they operate on and here is how they would change these knowledge structures under
such and such circumstances.
Attention with regard to research on learning within cognitive psychology for the last decades has, to the
extent it has existed, been very much focused on cognitive, conceptual learning. Shuell (1986, pp. 415–418)
has summarized features characterizing contemporary cognitive learning research. He distinguishes five
features concerning the impact of cognitive psychology on learning theory:
a) A view of learning as an active, constructive process
b) The presence of higher-level processes in learning
c) The cumulative nature of learning and the corresponding role played by prior knowledge
d) Concern for the way knowledge is represented and organized in the memory
e) Concern for analysing learning tasks and performance in terms of cognitive processes that are involved.
The dominant contemporary cognitive approach to learning research conceives of learning as an “active,
constructive and goal-oriented process that is dependent upon the mental activities of the learner” (Shuell,
1986, p. 415). It is evident that learning is seen as dependent on some mechanisms 
within
an individual
subject. Shuell (1986) concludes that cognitive learning “focuses on the way in which people acquire new
knowledge and skills and the way in which existing knowledge and skills are modified”. He further argues
that almost all conceptions of learning involve the following three criteria: (a) a change in an individual’s
behaviour or ability to do something, (b) a stipulation that this change must result from some sort of practice
or experience and (c) a stipulation that the change is an enduring one (p. 412).
In Shuell’s summary the question of 
what
is changing in learning is defined in the first paragraph (a): it is
the “individual’s behaviour or ability to do something”. The second feature again refers to the conditions
under which this change may occur; it excludes changes on the basis of how these changes come about.
Changes in (a) caused by, for example, maturing are not counted as learning.
It is also evident that thinking is stressed in this view of learning; the “active, constructive and goal-
oriented” processes are discussed in terms of “mental activities”. The following features are regarded as
typical of this research (Shuell, 1986):
a) The role of meta-cognitive processes such as planning and setting goals and subgoals…
b) The active selection of stimuli…
c) The attempt by learners to organize the material they are learning, even when no obvious basis of
organization is present in the materials being learned…[and]
5. OBJECT OF ANALYSIS
107

---

d) The generation of appropriate responses and the use of various learning strategies.
In several respects the general characteristics of the cognitivist approach are also reflected in an article by
Robert Glaser from 1987. He claims that learning is seen as “conceptual shifts from surface-level problem
representations to deep structure representations” (p. 397). In these conceptual shifts, information is
restructured. Learning is thus a restructuring activity which is accompanied by “an accumulation of new
facts, rules and procedures” which also involve “acquiring different relations among concepts” (ibid.). The
result of learning is that the individual gets a “large collection of knowledge structures that enable
representation of relationships in a problem” (ibid., p. 398). Further, in learning “humans …build up an
extremely large store of structured knowledge” (ibid., p. 400). In studying learning one must therefore
analyse “the processes by which learners integrate and organize information” and “how cognitive structures
are modified and combined” and further how students “learn …to construct interpretations of situations that
occur in the course of learning” (ibid., p. 398). If we follow Robert Glaser, we must, in order to understand
what learning means more precisely within the information processing approach, look at how new facts are
accumulated, what it means to build up a large store of knowledge, and how learners process information.
We can see that learners’ activities are emphasized strongly. Learning is thus in the hands of the learner.
Gunstone and Northfield (1994, p. 545) have expressed this position clearly:
It is the pupil or student teacher who must first