Beyond Constructivism

JONATHAN F. OSBORNE
Kings College, University of London, Cornwall House, Waterloo Road, London SE1
8WA, UK; e-rnail: J.Osborne@kcl.ac.uk

This article offers a critique of constructivism in science education which is an attempt to define and identify not only the weaknesses, but the successes of constructivism. Its success has been to generate a significant body of empirical data which
has contributed to our knowledge and understanding of difficulties in the learning of
science. That knowledge has also enabled the development of some innovative
methodology for the pedagogy of science and a greater awareness of the learner.
However, as a theoretical referent, it suffers from a flawed instrumental epistemology which is a misrepresentation of the views and practice of science and scientists.
Further it has confused the manner in which new knowledge is made with the manner in which old knowledge is learned, assuming that the two are one and the same
thing. The result of these failings is that it offers no guidance on adjudication between theories, the organization and sequencing of content within the science curriculum and rejects any value for didacticism. The failure of constructivism to
recognize its own limitations has lead to it enjoying a hegemony in the research
community which is undeserved. Instead, it is proposed that an alternative of modest
realism offers, not only a better representation of the practice of science, but additionally some value in determining issues of pedagogy. 0 1996 John Wiley & Sons,
Inc.

INTRODUCTION

Constructivism in science education has its roots in a reaction against two features dominating science curriculum reforms in the 1960s and 1970s-an epistemology based on naive empiricism (Harris & Taylor, 1983) and a developmental
stage model of cognitive growth which was interpreted as implying deterministic
limitations to childrens capabilities. Driver and Easley ( 1978) initiated the reaction
to these schools of thought when they argued that achievement in science depends
to a greater extent upon specific abilities and prior experience than general levels
of cognitive functioning. From here, research based on this premise developed
rapidly, as did attempts to evolve a theoretical base for the empirical findings, to
the point where the growth of publications in this field has become almost exponential (Duit, 1993). Such is the dominance of this school of thought that it has
been contended that it offers a new paradigm for science education (Tobin et al.,
Science Education 80(1): 53-82 (1996)
0 1996 John Wiley & Sons, Inc.

CCC 0036-8326/96/010053-30

54 OSBORNE

1994) and it has acquired implicitly the status of a meta- or grand theory
(Lyotard, 1984).
In developing its case, constructivism has focused very strongly on the resilience of the learners beliefs and the social construction of reality. When such
features are in focus, there are other features which are blurred and out of focus, if
not out of the picture altogether, and the consequence of the concentration on
these issues has led to serious epistemological flaws in these constructivists conceptions of science, that is, the manner in which new knowledge is made.
Furthermore, to the extent that the personal and social has been given priority
over the obdurateness of the natural world, it is contended that it fails to distinguish between real and theoretical entities. The result is an instrumentalist epistemology and a misrepresentation of the nature of science through an overemphasis
of the construction of the concepts, either personally or through discourse, and a
failure to elaborate any methodology of theory adjudication. Notions of truth have
simply been replaced by the concept of viability, and the failure to examine how
one idea could be considered to be more viable than another is at heart a denial
of objectivity and the rationality of science. Additionally, constructivist pedagogy
often makes a fallacious connection between the manner in which new scientific
knowledge is created and the manner in which existing scientific knowledge is
learned.
At heart, there is a failure by constructivism to recognize its limits, that is, to
identify it strengths and its weaknesses. As a method it has had considerable success in its critique of didacticism. Moreover, the types of alternative learning
strategies it offers, and has generated, have made an important contribution to our
understanding of the learner and learning (Solomon, 1994). The large body of empirical data that has been generated by work in this research paradigm has been
seminal in improving teachers knowledge and understanding of childrens scientific thinking, its origins and its development. However, as a referent (Tobin &
Tippins, 1993), it suffers from flaws which will always restrict its potential and any
claims to universality.
WHY EPISTEMOLOGY MATTERS

Science earns its place on the curriculum because there is a cultural commitment
to the value of the knowledge and the practices by which this body of ideas has been
derived. Hence, any consideration of the theoretical base of science education and its
practical implementation must start by attempting to resolve the aims and intentions
of this cultural practice. A variety of useful attempts have been offered. For instance,
Ogborn (1988) has proposed that science offers the answers to five distinct questions
which are all important aspects of science education:
0
0
0
0
0

What do we know? (The ontological question.)

How do we know? (The epistemological question.)
Why does it happen? (The causal question.)
What we can do with our knowledge? (The technological question.)
How we can communicate these ideas? (The communicative questions.)

BEYOND CONSTRUCTIVISM 55

From a different perspective, Hodson (1990) distinguishes three dimensions to an education in science- learning science, learning to do science, and learning about science. Millar (1995) argues that the aims of science education are to learn something
about the body of scientific knowledge, something about the processes by which that
knowledge is generated, and something about the manner in which knowledge is socially constructed. All of these perspectives have one fundamental issue in common-that knowledge and understanding of the epistemology of science is an
essential aspect of any education in science, and any approach which neglects a consideration of it is incomplete and epistemologically thin. Courses that fail to consider
how scientists come to know are running the risk of developing students who do not
acknowledge the scientists views as rational (Duschl, 1990). Some evidence of this
happening comes from the increasing prevalence of antiscientific beliefs and attitudes
accompanying the rise in post-material political values in modem youth (Inglehart,
1990).
Thus, a central thesis of the argument presented here is that epistemology does
rnatter-because the answer to the question how we know is an important aspect
of the account of science and the evidence that we present for our ontological commitments. Science without a serious consideration of this question would leave the
learner in the position portrayed by Horton in his seminal discussion of the nature of
African traditional beliefs where he concludes that:
The ground for accepting the models proposed by the scientist is often no different
from the young African villagers ground for accepting the models propounded by
one of his elders. In both cases the propounders are deferred to as the accredited
agents of tradition. . . . For all the apparent up-to-dateness of the content of his
world-view, the modem Western layman is rarely more open or scientific in his
outlook than is the traditional African villager. (Horton, 197 1, p. 262)

To transcend that position, an explicit treatment of the procedural basis of Western

scientific knowledge is essential in science education. This point is also emphasized by Duschl (1990), who contends that to paint a complete picture of science,
a science curriculum should address not only what is known by science, but also
how science has come to arrive at such knowledge. Thus, the emphasis on declarative and propositional knowledge at the expense of the procedural aspect of science, of which the British National Curriculum is a typical example, needs to be
challenged.
Furthermore, such a view requires that teachers have some understanding of the
epistemology of science-that is, the nature of the subject that they present to
children (Koulaidis & Ogborn, 1995; Shulman, 1986). While any treatment of the
nature of science is problematic, as the writings of Popper, Kuhn, Feyerabend, and
Lakatos show that there is no agreement that there is one methodological base to
science (Laudan et al., 1986), this does not mean it is correct to assume that it is
an irrational practice that has no methods. Science has a range and variety of
methods and it is important that science education generates some understanding
of how the practices of science legitimizes its knowledge claims. For, if one of the
aspirations of science education is to develop critical thinkers able to understand
and examine the myriad claims made on a scientific or pseudo-scientific basis,

56 OSBORNE
children will need to acquire a good set of intellectual tools. One of these is an understanding of how the processes deployed by scientists to ensure their reliability
and validity of their knowledge. Thus, concepts of fair testing, the identification
and control of variables, the generation of sound hypotheses, the recognition and
measurement of sources of error, the criteria by which one theory is distinguished
from another, repetition of experimental determination, and the use of averaging
are all essential ideas that enable scientific knowledge to be distinguished from
pseudo-science. A science education that omits discussion of these aspects or presents a flawed epistemology, is the intellectual equivalent of giving a child a hammer without the nails.
The view that an understanding of the epistemology of science is an important
component of science education is based firmly in the belief that:
Science effects styles of thinking and reasoning in society; it influences the acceptable modes of rational and reasonable debate; it contributes to public intelligence by
instilling a concern for evidence and its rational and open appraisal. For all its faults,
the scientific tradition has promoted rationality, critical thinking and objectivity. It
instills a concern for evidence, and for having ideas judged not by personal or social
interest, but by how the world is. (Matthews, 1995, p. 2)

Put simply, the justification of the place and time for science education on the curriculum lies in the claim that the methods of science and scientists produce reliable
knowledge, that is, that its epistemology is superior to that of astrologers and even
sociologists. Thus, any account of science requires a clear and well-defined understanding of the nature of the knowledge claims that science makes.

CONSTRUCTIVIST EPISTEMOLOGY
What then can be said of constructivist epistemology? The two most clearly
elaborated epistemological positions of constructivism are those of the radical constructivist and the social constructivist, both of which will be examined. Their
ideas need to be contrasted with those of traditional epistemology, which is based
on the view that knowledge is a success term. Successful in the sense that A is able
to make a claim p about the world because (a) p is a true statement about the
world; (b) A believes p; and (c) A has a tethering reason for believing in p. In this
manner, knowledge is to be distinguished from opinion and the requirement for
reason as the path from evidence to knowledge is the foundation of the rationality
of science.
Radical constructivism is a deliberate and public attempt to move away from such
an epistemological base. For instance, Glaserfeld (1993) argues that radical constructivism is postepistemological and an attempt to cut loose from the philosophical tradition that knowledge has to be a representation of reality. Instead, truth is replaced
by the notion of viability. Rightly, Glaserfeld recognizes that this is a stark step.
Essentially for him, knowledge is always the result of a constructive activity, and
therefore cannot be transferred to a passive receiver. Thus knowledge exists only in
the mind of cognizing beings, where it is construed, and is not to be found, for instance, in books, texts, or other traditional repositories of human knowledge which

BEYOND CONSTRUCTIVISM 57
simply represent symbols open to a range of interpretation. Hence there is no knowledge without a knower.
This is not to say that radical constructivist are solipsists. For, as Tobin argues:
Constructivism is a set of beliefs about knowledge that begins with the assumption
that reality exists but cannot be known as a set of truths. (Tobin et al., 1994, p. 47)

At the heart of many radical constructivist writings seems to be a desire to escape notions of absolute truth, which are accurate or even partial representations of the real
world. For instance, Tobin suggests that:
We construct a model of gravity that is viable in that the model fits experience, but
no matter how elegant, that model cannot claim to be an absolute truth. (Tobin &
Tippins, 1993, p. 2)

Furthermore, he continues that we cannot tell whether our model for gravity gets
closer to an absolute reality. Instead these authors epistemology requires only that
knowledge is viable because of its coherence with other understandings and it fits
with experience. Thus truth becomes an act of faith, and knowledge has pragmatic
validity in that truth is what works. The only requirements should be that it fits with
experience and is coherent and subject to empirical verification both personally and
in the wider society.
There are a number of epistemological problems with this position. First, it attaches itself to the bivalence principle, knowledge is either true or false, and since
we can never know absolute truth, all knowledge is only subjective, provisional
and uncertain. For the radical constructivist, there is no halfway house, no asymptotic approach to any zenith of human understanding and no admission that what
we know improves and increases. Thus the unattainable best is used to drive out
the attainable better (Ogborn, 1994), and instead, we are simply abandoned on the
ocean of experience condemned to search for knowledge that is viable. Better
ideas may be developed in the sense that they are more viable, and fit with experience, but we are not necessarily nearer an objective representation of the physical
world.
The position of the radical constructivist is essentially that of the instrumentalist.
Like the latter, theories are merely portrayed as convenient devices for describing
phenomena and connecting one set of events with another. They are best understood
as useful fictions which bear no necessary relationship with reality and one idea is
only considered better than another only if it is a more useful calculating and predictive device. Such a position is most aptly summarized by Osianders comment in the
17th century on the theories of Copernicus:
For these hypotheses need not be true or even probable; if they provide a calculus
consistent with the observations that alone is sufficient. (Rosen, 1962, p. 125)
However, the problem for the instrumentalist is that these theoretical fictions have an
unfortunate habit of predicting objects, which due to the advances in technology and
instrumentation can increasingly be seen or sensed. Thus, KekulCs benzene ring

58 OSBORNE
structures are revealed by the electron microscope, Maxwells dynamical field structures are made tangible by Hertz, Einsteins predictions of the stimulated emission of
radiation lead to the development of a technological artifact which is a daily, commonplace occurrence.
The constructivist position is essentially similar to that of philosophers like Rorty
(1989), an influential postmodernist, who argues that reality is indifferent to our descriptions of it and sees language not merely as a tool for expression and representation but the means by which the human self is created. For Rorty, arguments about
truth and realism are irrelevant as there is no metaphysical position which transcends
the limitation of our own vocabulary and would allow us to adjudicate between competing knowledge claims. Therefore, the interesting question becomes not How do
you know? but Why do you talk that way? Rorty is correct in stating that reality is
indifferent to our description of it and there are often multiple interpretations for the
same event or observation, but in science, our language is bound by reality. For science does have a well established methodology and acts of reference for deciding between competing theoretical descriptions and rejecting those which are clearly
fallible, incomplete, or simply false. Furthermore, the content of the scientific knowledge that forms the substance of science education is not some transient, uncertain
representation, where one theory is replaced by another that is incommensurable, for:
If science is all that ephemeral, if Newtons second law must be classed as a temporary codex, it is somewhat surprising that science has been as useful as it has in human understanding. (Weinberg, 1968, p. 74)

But how is the child to judge that the scientists description is more viable? How is
the child to know that there are entities for which we have well-established arguments for their existence and reliable theories that have superior explanatory power
than those of common-sense reasoning? Constructivism singularly fails to elaborate
any mechanism by which one theory can be considered more viable than another.
Glaserfeld (1991) asks merely that we try to develop a theory which offers a relatively coherent explanation and that such theories are tested in the experiential
world where they either do or do not do what they are claimed to do. The problem
is that many theories, including Ptolemaic astronomy, would meet such criteria.
Tobin, at least, implicitly (see previous quotation) recognizes that some theories are
more elegant than others but fails to elaborate any criteria by which such judgments could, or should, be made. In defending the notion of viability, he argues:

. . . that knowledge must be viable not only personally, but also in the social contexts in which actions do occur. Viability is therefore determined with respect to
those actions which facilitate the attainment of goals in the social contexts of action.
(Tobin & Tippins, 1993, p. 3)
However, all this argument does is extend the notion that ideas must be personally
viable to the idea that they must be viable within a social context. Yet again, it fails to
elaborate any criteria that determine the concept of viability.
A full consideration of how we adjudicate between theories would require an acknowledgment that our language and our ideas are constrained by reality, to real re-

BEYOND CONSTRUCTIVISM 59

ferents that do exist, and thus, while there may be no ultimate check that any construction is right, there are checks that one construction is better than another. The
failure to recognize this important practice of science leads implicitly toward a relativist ontology where viability is equated with validity so that any viable theory is
considered worthy of consideration.
Consequently, a fundamental epistemic problem for constructivism is a failure to
consider the issue of theory adjudication and articulate a mechanism to establish how
one theory may be considered to be better than another. Science becomes a set of
meanings and concepts which evolve from socially negotiated understandings and
not from a process of examining whether such inventions are supported when tested
against the real world. Thus, the childs common conception that the sun moves
round the earth clearly satisfies the radical constructivists epistemological criteria
that it is both viable and fits with experience. Moreover, such understanding may be a
socially negotiated product of the classroom or playground but on what basis is it to
be found wanting? For the scientist, theories are successful because they offer a
range of explanation, non- ad hocness, consistency with empirical evidence, and logical consistency which gives them explanatory force that inductive generalizations
lack (Samarapungvan, 1992). Theories that offer parsimony and that accrue the epistemological status acquired by earlier theories are also preferred. The point here is
not to elaborate all the criteria by which theories are judged but simply to show that
such explicit criteria do exist and are deployed by scientists in making judgments
about their knowledge claims.
Such authors seem unaware of the interest and debate in the philosophy of science
which has given much prominence to the criteria for theory selection and the importance of this issue to the epistemology of science (Giere, 1984; Laudan et al., 1986).
In part, the failure to explicate the methods and criteria of science may be because of
the attack from the sociologists of science who argue that:

. . . the conviction that a body of practices widely regarded by outsiders as well organised, logical and coherent, in fact consists of a disordered array of observations
with which scientist struggle to produce order. (Latour & Woolgar, 1986, p. 36)
Thus, scientific activity is not about nature, but is a fierce fight to construct reality,
and consequently:

. . . science is a form of fiction or discourse like any other, one effect of which is
the truth effect, which (like all literary effects) arises from textual characteristics,
such as the tense of verbs, the structure of enunciation, modalities and so on. (Latour
& Woolgar, 1986, p. 184)
Clearly, the view projected here is that science is a cultural product, distinguishable
only by its form and not its methods. These interpretations of science (Collins, 1985;
Latour, 1985; Pickering, 1984) would deny that scientists deploy a set of criteria in a
rational and consistent manner to evaluate their ideas. In essence, their argument is
that the claims of science to follow explicit prespecified procedural rules, to possess
explicit criteria for correctly recognizing cases, to have complete justifications at
each step, and to have clear predetermined goals are simply not tenable. Anything

60 OSBORNE
more than a superficial examination of scientific practices shows that all of these features are never present, thus the knowledge claims associated with science lack the
rationality and the truth they seek to assert. Science thus creates the illusion of certainty.
However, a closer examination of their argument shows that it lacks credibility.
Scientists clearly do have criteria for judging specific evidence for and against any
given theoretical position. The fact that all criteria are not always applied or used in a
consistent manner is not a justification for questioning the whole edifice of science.
Such an action is rather like a composer discarding a whole symphony because of a
few bad notes. Scientific thinking is a social construct and scientists do work in communities of practice with established norms. Radical ideas are never easily accepted
and occasionally can result in a collapse of the moral order that constitutes the practice of science. But the lack of consistency in the application of rules to its epistemic
practices does not mean that the scientists epistemology lacks any rules. There are
clear rational arguments why general relativity is a better description of nature than
Newtons cosmos or why Mendelian genetics is to be preferred as an explanation of
inherited characteristics compared to Lamarkian theory. Unfortunately, constructivist
epistemology fails to acknowledge these claims and, in so doing, undermines the
very core of what makes scientific knowledge different from other forms of knowledge. For, if the epistemology it argues for is a fallacious representation of science, a
form of discourse like any other, why should the knowledge and products of scientific discourse be given any authority? Science leaves itself with no defense against
the argument that it is a self-sustained delusion vehemently mounted by the historian
Bernard in Stoppards play Arcadia:
A great philosopher is an urgent need. Theres no rush for Isaac Newton. We were
quite happy with Aristotles cosmos. Personally, I preferred it. Fifty-five crystal
spheres geared to Gods crankshaft is my idea of satisfying universe. I cant think of
anything more trivial than the speed of light. Quarks, quasars-big bangs, black
holes-who gives a damn? How did you people con us out of all that status? All that
money? And why are you so pleased with yourselves? (Stoppard, 1993)

SOCIAL CONSTRUCTIVISM
The social constructivist perspective portrays science education as a process of enculturation in which the aspirant members of a culture learn from their tutors and
that novices are introduced to a community of knowledge through discourse in the
context of relevant task (Driver et al., 1994). What might be termed the strong form
of this position argues that learning is a form of cultural apprenticeship and that cognition is situated in specific contexts (Lave, 1988; Brown et al. 1989); that is, that
learning cannot be separated from the context of learning. Now in that learning to be
a scientist, or to understand what a scientist does, is a process of cognitive apprenticeship there is no quarrel. If as Harrk (1986) argues, science represents the epitomy of rational cooperation set within a strict moral order this is best appreciated
from working within the discipline or a context that replicates it as near as possible.
That schools often fail to do this is accepted and the consequence is, as Brown et al.

BEYOND CONSTRUCTIVISM 61
(1989) argue, that science education has become a set of ersatz activities which are
not those of scientists. What is missing from their account is any understanding of the
rules and practices that guide the scientific community. They present an opaque epistemology where knowledge indexes the situation in which it arises and is used . . .
which becomes coded by and connected to the activity and environment in which it is
developed. Epistemological considerations of how it is coded or connected to the
activity are never discussed.
Driver et al. (1994a) do at least define what they see as the core ontological commitments associated with the practice of science. For them:

. . . scientific knowledge is symbolic in nature and socially negotiated. The objects

of science are not the phenomena of nature but constructs that are advanced by the
scientific community to interpret nature. (Driver et al., 1994a, p. 5 )
Thus, the things of science are no longer entities that populate a real world but the
inventions of a scientific discourse which are imposed on the phenomena. Hence it
is not nature which limits the scientific imagination but simply human capabilities
and their cultural and conceptual tools which reflect the cumulative wisdom of the
culture. Basically the confusion that occurs here is to conflate ideas and objects, that
is, to fail to discriminate between the objects of the discourse with the statements of
the discourse. The two are not the same, one exists and can be experienced, often
with the aid of instrumentation, the other is a socially negotiated construct.
The result is that the social constructivist would portray the world as being full of
peculiar symbolic entities such as atoms, electrons, ions, fields, genes, and chromosomes. These are organized by such ideas as evolution or the nuclear model of the
atom, but essentially there is nothing special about the truth of science. Science is
compared to learning a foreign language so that
If some foreign languages are more difficult for you to learn than others, that is
mainly because they are less like the language you already know, or the experiences
they represent are less familiar. Science has its own distinctive genres, its thematic
formations, its practical skills. In their forms they are no more intellectually complex
or difficult than those of any other subject. They are less familiar, less like what we
already used to. (Lemke, 1990, p. 139)

Lemke extends this position further to argue that every statement of fact is a statement that is useful for some purposes to act as if something were true. Theories and
hypotheses are just that, statements useful for some purposes. Once again, this is
the instrumentalist vision of science, and again the mistake made here is not the representation of science as a human product which is socially constructed, but to confuse the manner in which knowledge is made with the objects of knowledge
themselves. They are not one and the same thing. Lemkes argument would appear to
be that learning science is no different from the problems faced in learning English or
history, and is distinguishable only by the form and content of the discourse and not
by the manner in which that knowledge is produced. Thus, learning science is simply
a matter of learning to talk the right language. Furthermore, the hallmark that distinguishes these two forms of knowledge is not what they describe, but how they

62 OSBORNE

describe the world. Essentially for the social constructivist, scientific theory does not
describe the world, but on the contrary, constitutes it.
Close examination of some of the sample scripts offered by Driver et al. (1994a) as
exemplars of the construction of scientific knowledge shows the nature of this problem. The authors use these to argue that they illustrate the discursive nature of scientific knowledge from which children have to make personal sense. The teacher is
presented as a guide mediating and negotiating between the childrens everyday
world and that of science. Yet never once within the several dialogues does the
teacher address important epistemological issues and ask of the student How do you
know?; How would you justify that view?; or Why do you believe you are
right? That is, the issue of evidence and the act of reference are omitted, and instead, the function of the teacher is to negotiate with, and persuade the student
toward a scientific way of seeing through discourse. Yet in what way is such practice
to be distinguished epistemically from Hortons African child who is imbued with his
forms of thought through the oral culture of his tribe?
For the realist, even the modest realist, such a position is unacceptable.
Scientists act on the world, they intervene, manipulate, and measure. Their language and discourse acts are limited, not by their imaginations or cultural conditioning, but by the evidence they gather from such actions on the world. For we
may think what we like, but we cannot do what we like. Thus, nature always constrains our discourse:
The world contains natural substances with peculiar properties. These obey laws
which fix the (infinite) range of possible ways in which the natural substances behave in given circumstances. Some of the possible ways in which these substances
behave are revealed in the quite contingent course of the actual history of the world
without any.human intervention. Other possible ways are revealed when humans intervene contingently in natural processes, for example, in the laboratory. The human
intervention is such that, with the growth of theory (constructed by humans), we are
able to manufacture, in the laboratory, new materials that would not have emerged in
the history of the world without human intervention. We might say that natures substances have underlying dispositions, or powers, some of which will be exhibited
while others may not be triggered at all. . . . This is a quite traditional metaphysical picture which I think accommodates much of what constructivists wish to say
about science without indulging in radical worldmaking. (Nola, 1995a)

The social constructivist stands in sharp contrast to this position. For instance
Latours third rule of method (Latour, 1985) argues that since the settlement of a
controversy is the cause of Natures representation, not the consequence, we can
never use the outcome-Nature-to
explain how and why a controversy has been
settled (Latour, 1985). Thus, science is portrayed as a way of world-making rather
than being made by the world.
Ultimately, the social constructivist who sees knowledge as primarily fashioned
and made by human discourse has no defense to the contention that science is an irrational enterprise where persuasion and force of argument are the main criteria by
which theories are to be judged. The realists defense of such accusations lies in the
fact that their discourse acts refer and their veracity can be checked by careful replication of the knowledge claims.

BEYOND CONSTRUCTIVISM 63

CONSTRUCTIVIST PEDAGOGY
How is constructivist theory applied to the process of teaching and learning?
Tobin and Tippins (1993) make a useful distinction when they ask whether constructivism is a method or a referent. As a method, its value is for the production of
effective pedagogic practices; as a referent, it offers a general and wide-ranging
theory of learning. Tobin and Tippins see the former as a devaluation of the power
of constructivism. However, it is perhaps as a method that constructivism has had
the most value in education. The strong message of this body of research has been
to expose the difficulty large numbers of children experience in internalizing the
explanatory models of science and applying them correctly. Such data has inevitably raised the dilemma of how to respond to this research evidence. The constructivist response has been twofold. First, there has been an attack on the notion
that children are atheoretical (Carey, 1985) and the notion that teaching is a process
of filling empty minds or tabula rasa (Driver, 1989; Driver & Oldham, 1985;
Gilbert et al., 1982; Hewson, 1981). Solomon (1994) has identified the greatest
success of constructivism when she argues that redescribing a network of related
things by a new vocabulary creates a pattern of language which itself constituted a
new theory. The shift in describing pupil errors from mistakes, of no theoretical
interest, to misconceptions, or more often alternative frameworks, changed the
commonplace and unremarkable to something significant. Much of the initial appeal of this research lies in the way its findings give significance to the everyday
experience of the teacher. Therefore, constructivism has encouraged teachers and
curriculum developers to alter their perceptions of children from individuals who
are irrational and unknowing to cognizant beings who have well-developed theories. Formal elicitation of this knowledge as proposed by Driver and Oldham
(1985) now becomes important for two purposes-to encourage the child to clarify
and articulate their own understanding; and as a process of formative assessment
by the teacher to ascertain the teaching and learning needs of their students.
Teaching becomes a process where understanding and knowledge are no longer
transmitted but negotiated (Sainsbury, 1992) through a process of cognitive apprenticeship (Brown et al., 1989).
Constructivists then rightly turned their attention to the learner arguing correctly that he or she is responsible for their own learning (Osborne & Wittrock,
1985; Novak & Gowin, 1984; Pope, 1985; White, 1988). Here, after all, in the
learners mind, was where new meanings had to be formulated and understood.
This could only be achieved if he or she was an active participant in the learning
process. Hence, their pedagogy has concerned itself with a program of activities
from which knowledge can be generated or acquired. These tasks focus on the individual subject asking them to articulate and use their reasoning in a set of structured exercises often making use of groups. Many of the earlier schemes made
good use of group discussion and poster-making for such purposes (Nussbaum &
Novak, 1981; Cosgrove et al., 1984; CLIS Project, 1987) and such processes enable the social construction of meaning which is an essential aspect of science education. Further work by many researchers in the field has led to the production of
a wide range of structured techniques that require the active participation of the
student, and one recent book (Baird & Northfield, 1992) gives details of more

64 OSBORNE

than 80 such techniques. Notable among these structured strategies are predictobserve-explain sequences (White, 1988; White & Gunstone, 1992); discussion
of instances of physical phenomena (White & Gunstone, 1992); concept mapping
(Novak & Gowin, 1984); word association (Shavelson, 1974); and active reading
techniques commonly called DARTS (directed activities related to text) (Davies &
Greene, 1984). However, it should be noted that neither of the latter two strategies
were the product of an explicitly constructivist approach to pedagogy, and currently, there is only limited research evidence for the value of any of these approaches. One notable exception is Hortons (1992) recent meta-analysis of 18
studies of concept mapping that met strict criteria of well-defined experimental
models with controls for which he found 16 that produced positive learning gains
for the experimental groups. For the sake of establishing their case, more empirical studies of this nature are essential to justify the value of constructivist approaches to teaching and learning.
Moreover, the other great contribution of empirical research conducted from this
paradigm has been the provision of a very broad and detailed knowledge of childrens thinking and the problems that they encounter with scientific ideas. Making
Sense of Secondary Science, by Driver et al. (1994b), is an excellent summary of
work that has provided useful insights into the common-sense origins of childrens
thinking and the stages of its development within a domain which have guided both
curriculum developers and teachers.
For Tobin, however, the appeal of constructivism is not as a method, nor in the empirical research findings, but its wider potential as a referent:
To be a viable theory of knowing, constructivism must have explanatory power in all
situations in which knowledge is constructed. . . . Thus constructivism can be used
as a referent to analyse the learning potential of any situation. (Tobin et al., 1994,
P.47)

Such claims are tinged with the view that constructivism is a grand theory, applicable to all instances and not simply a view of learning that has partial value.
However, as a referent, the limitations of constructivism are indicated by a number of omissions. First, there is the failure to use the theory, as opposed to the empirical research findings, to predict or discuss what it might imply for the content
and process of science education, and to derive any detailed propositions which
could in principle be empirically tested. For instance, the term metacognition is
used to describe the thinking generated by active learning. Yet the lack of any theoretical treatment that adequately describes such activity inevitably leads to more
unanswered questions. Are all pupils capable of metacognitive activity or just
some? Is there a critical age below which children cannot be metacognitive? Is all
such activity beneficial? Most published work on the topic (White & Mitchell,
1994; Gunstone, 1991) very much takes the pragmatic view that metacognition is
inherently a good thing and fails to make Browns (1987) important distinction
between knowledge of cognition and the regulation of cognition, or to note
her point that most workers in the field have found little evidence of the former
before adolescence. So, while the interest in metacognition and the notion that the
pupil should be active is laudable, such omissions make it questionable whether

BEYOND CONSTRUCTIVISM 65

constructivist theory is sufficiently comprehensive to be adopted as a universal

referent.
Second, the question of content inevitably re-emerges. The question that needs
to be asked is: Metacognitive about what? To the teacher and the curriculum developer such decisions are the bread and butter of their daily lives as they attempt
to order content to form a coherent introduction to science, and to select material
which stimulates but is not overdemanding. As a referent, constructivist accounts
of science education have nothing to say other than the familiar Ausubelian refrain. The problem of the determination of content is compounded by the fact that
the theoretical base in Ausubelian psychology and that of Kelly (1955) has never
been developed into a model which would enable predictions to be made of what
material will or will not be accessible to learners. The much quoted statement of
Ausubel (1968) on the significance of educational psychology is essentially nothing more than a statement of good common sense, and the success of constructivism has come from reminding teachers that children do have ideas and theories
and that their thinking is the foundation on which new meanings must be formulated.
A more serious criticism of the constructivist theory is that it provides no welldefined mechanism by which the individual can develop new constructs with which
to see the world. From whence come the ideas with which the individual is to interpret their sensory perceptions? Where, for instance, is the role of analogy and
metaphor which are the vehicles for extending our thinking and ideas and reorganizing our internal symbolic representations? For example, to observe Brownian motion
in a smoke cell, a student has to be provided by a teacher with a construct that will
enable them to make sense of their perceptions prior to observation. This can only be
done through the use of a taught analogy or comparison. Without this, the common
experience is that the students attention is needlessly focused on other elements in
the microscope. Moreover, reformulation of sensation will only occur by reflection
and reorganization of the representations, something which Cosgrove et al. acknowledge when they state:
We are increasingly aware that conceptual change takes time and that it is important
that counterintuitive ideas are considered at regular intervals over a period of time.
(Cosgrove et al., 1984, p. 254)

What is omitted from the constructivist writings in science education is comprehensive account of those processes which would help the individual to generate new
ideas and concepts, for reinterpreting experience, and transcending common-sense
reasoning. Intervention studies have predominantly focused on conceptual conflict
(Cosgrove et al., 1984; Gunstone & White, 1981; Hewson & Hewson, 1984; Stavy &
Berkovitz, 1980; Thorley & Treagust, 1986) which have shown has only limited effectiveness (Cosgrove et al., 1984; Rowel1 & Dawson, 1983). In addition, school
practical apparatus is notoriously unreliable as a method of providing direct experience of phenomena and is not a good vehicle for developing theory (Osborne, 1990;
Woolnough & Allsop, 1986). Driver herself does in fact come near to confronting
this very issue when she states:

66 OSBORNE
The theoretical models and scientific conventions will not be discovered by children through their practical work. They need to be presented. Guidance is then
needed to help children assimilate their practical experiences into what is possibly a
new way of thinking about them. (Driver, 1983, p. 9)
But, in her later seminal study on a constructivist view toward curriculum development (Driver & Oldham, 1985), this matter is avoided as we are told that we can
specify the experiences which students should be exposed to but are given no principles on which to make such judgments other than knowing where students are
starting from and a reliance on the teachers intuitive knowledge of classroom realities. The curriculum process itself contains an evaluation phase where pupils test a
range of ideas including the scientific one if they have suggested it. But what if
they have not? (On this issue the authors remain silent.) In a more recent position paper from the Leeds constructivists (Driver et al., 1994a), only once in all the exemplars used to illustrate their conception of the process of learning does the teacher
assert some scientific knowledge and, even then, without justification. Therefore,
constructivists, and science educators in general, must recognize that there is a role
for telling, showing, and seeing as methods which enable the construction of new
knowledge.
Faced with the theoretical problem of the determining content, one can only conclude that constructivism has eschewed the issue. Instead, workers in the field have
resorted to an individualistic approach, arguing that opportunities must be provided
for the learner to externalize their understanding and this must then be challenged by
critical incidents which generate conceptual conflict (Driver & Oldham, 1985), in
such a manner that the new ideas are plausible, intelligible and fruitful (Hewson &
Hewson, 1984). -While such strategies are a valuable contribution to the teachers
repertoire, they are not a complete account of the teaching process.
For instance, as a comprehensive referent, any perspective requires a consideration of whether the student has the cognitive tools to undertake the manipulation of
symbols and whether there is age-related development in such facilities. For some
science educators (Adey & Shayer, 1994; Shayer & Adey, 1981), sequencing of
content is possibly the most important issue. However, the instrumental epistemology of the radical or social constructivist fails to recognize any distinctions in the
complexity and difficulty of theoretical descriptions of real objects, or that some of
the idealizations of science describe concepts which are not accessible to sensation or experience, for example, energy, but have to be formulated through a
process of reflective abstraction. While the notion that knowledge is personally
constructed is at one level a trivial truism, the operational processes required for
the construction of knowledge are complex, poorly understood, and nontrivial. The
whole issue of conceptual complexity and whether there are developmental limitations on childrens capacity to undertake such reasoning is an issue that a constructivist theory ignores.
Finally, constructivist theory suffers from another confusion that questions its validity as a referent. The metaphor consistently underpinning constructivism is that of
knowledge as made, produced, generated by the individual (e.g., making sense).
For instance, Glaserfeld (1987) argues that

BEYOND CONSTRUCTIVISM 67
The cognitive organism tries to make sense of experience in order to better avoid
clashing with the worlds constraints. . . . Basically to have learned means to
have drawn conclusions from experience and act accordingly. (Glaserfeld, 1987,
p. 8)

Yet the emphasis on experience as the manner in which the individual learns is
strongly associated with a second metaphor of knowledge as found, lying around
out there to be discovered by the curious. The first metaphor is what the individual
has to do to learn and acquire new knowledge. Any individual when confronted
with a new phenomenon has to construct and internalize a meaningful representation for themselves-in such a manner is knowledge made by the individual.
The latter metaphor, which is ironically that of inductivism, is one perception of
how new scientific knowledge is produced but is rarely the manner in which old
knowledge is learned in the classroom. The lack of clarity and precision has resulted in a mixing of metaphors and the running together of two quite distinct
matters-the nature of scientific knowledge and the nature of learning (Ogborn,
1994). As Ogborn points out, there is no necessary connection between the epistemology of science and the methods by which it is learned, for it is like arguing
that the way to learn poetry is by writing poetry, a relationship which at best may
have only partial justification. This point has also been made by Hodson (1988),
Margetson (1982), and others.
However, Dana and Davis (1993) maintain a different view, arguing that the traditional objective view of knowledge leads to the belief that the process of teaching is
one of telling and the process of learning one of memorization and recall. Now while
this may be a true account of what has happened, and still happens in many societies,
it is not a necessary or logical connection. As one author has pointed out:
From the definition of objectivism as the search for truth, nothing follows about
learning discovered truths as the mere passive absorption of facts and their regurgitation. . . . It is quite possible to maintain objectivism, . . . and reject the view that
the teaching or learning of already discovered truths has to be conducted in such a
crudely didactic manner. (Nola, 1995b, p. 4)

Thus, the nature of science has no necessary bearing on the nature of teaching and
learning science, for the former is a philosophical issue to do with the nature of science while the latter is an educational issue to do with the best way of educating nonscientists about science.
Moreover, just as traditional, objective teachers have often failed to recognize the
need to structure learning so that it is active, the advocates of constructivist methods
of teaching have failed to recognize that there is a role for telling, showing and
demonstrating. Teachers, we are told, should negotiate, facilitate, coconstruct,
mediate, socialize, provide experiences, introduce, and make the cultural
tools of science available, but never ever will they tell. Clearly what is needed here
is the intellectual recognition that the commitment to a particular view of science
says nothing about the manner in which it is taught. Just as it is fallacious to infer
that objectivism implies didacticism, it is false to assume that a belief in constructivism implies that all knowledge must be negotiated. The rejection of the role of

68

OSBORNE

telling in pedagogy only confirms the suspicion that constructivists have difficulty
with the notion of objective knowledge-and that what each student constructs
needs only to make sense or be viable. That is, it is the act of constructing that is
important and not the product of construction for constraints on what constitutes scientific knowledge are not recognized.
Some evidence for the role and value of direct instruction can be found in the work
of Muthukrishna et al. (1993). They point to the fact that, in a number of studies that
address alternative frameworks directly, the success rate in changing childrens
ideas, which ranges from 28% to 69%,is limited, and argue that their research, using
an approach based on explicit instruction with a laser videodisc, which eliminated
over 90% of common alternative frameworks shows that the common hypothesis
that meaningful learning cannot result from explicit instruction may possibly be an
overstatement. While their work can be criticized on the basis that the domain of
earth sciences, in which this study took place, has little possibility of generating
many strongly held intuitive ideas, it does show that not all science education requires a singular approach.
In short, it would appear that there are significant problems for a paradigm which
seeks to be a referent for pedagogic practice. By emphasizing its successes and conveniently ignoring its weaknesses, constructivism has acquired the umbrella of an allembracing theory, when it is at its best, a partial contribution to science education.
This is not to argue that the process of education should be one of transmission, but
simply that it must enable children to acquire and understand the powerful constructs
and ideas of modem science and that any comprehensive theory of science education
should address the issues of content, concepts, sequencing, cognitive demand, and
the adjudication of competing theories.
What constructivism does have to offer is a useful focus on the learner as an active
participant in the learning process, a range of pedagogic practices which encourage
the child to be active, and a body of empirical research findings which are invaluable
for sensitizing teachers to the forms of scientific thought that children hold and their
possible stages of development. The latter is particularly important, for if teachers are
to change childrens understandings, it is important to be receptive to, and aware of,
the manner and origins of childrens thought. Put simply, to teach a child about science, not only does it help to know something about science, it helps to know something about the child. Research conducted from a constructivist paradigm in the past
decade has made a major contribution to the latter.
WHAT DO SCIENTISTS

DO?

The position taken here is that science education should attempt to communicate
not only the practice of science, but also the beliefs that are widely held in the scientific community, and one of the basic tenets is that scientific knowledge is different.
Scientists establish their knowledge claims through a program of experimentation,
verification, and peer review so that there exists a consensus which becomes their
touchstone of reliability. Therefore, the position of constructivist authors does not
represent the stance of most scientists whose views are more akin to that of Ziman
who argues that:

BEYOND CONSTRUCTIVISM 69
The objectivity of a well established science is thus comparable to that of a wellmade map drawn by a great company of surveyors who have worked over the same
ground along many different routes. (Ziman, 1979, p. 108)

And just as the child derives his or her well-founded confidence in the reality of
everyday things from the sensorimotor bodily experience of childhood, by acting and
intervening on the world, so does the scientist gain their confidence in the reality of
scientific entities which are initially theoretical speculations. As cells are regularly
manipulated under the microscope, or particles produced and detected in accelerators
and bubble chambers, our belief in the reality of these entities and our knowledge of
them grows.
However, scientists thoughts and ideas are constrained by the way the world
chooses to behave so that the methodological component of scientific epistemology
is one in which maps are created to store and represent scientific information, Such
maps are constructed by people who have worked over the same ground using many
different routes. Scientists confidence in the objectivity of their descriptions is
founded on the consensus of practice and ideas that emerges as the fruits of their
work and actions on the world. It is this link which distinguishes science from other
forms of knowledge. This is not to argue that science offers a moral exemplar to
other disciplines, but simply to state that it provides knowledge of an objective and
real world. Such knowledge is incomplete, curtailed, and limited but it grows through
the creative efforts of individuals who continually strive to formulate new hypotheses
about its nature and test these ideas experimentally. The ideas and theoretical entities
they derive are initially human constructions, often the product of a contingent cultural event in history and extremely tentative. Through experimentation, knowledge
of the adequacy and detail of these maps emerges, not always as in the case of
wave-particle duality, but often, as in the case of the electron, the AIDS virus, etc.
Just as the cartographers maps do not correspond with the terrain itself, so are the
scientists maps separate and distinct from the real entities that they describe.
However, in the same way that the former describe real features of the ground such
as its elevation, land use, and vegetation, so do the maps of the scientist describe real
features of the physical world. And just as the cartographer constructs new representations that more closely approximate to the physical terrain, so does the scientists
increasingly detailed descriptions give us a better knowledge of extant phenomena.
REALISM REVISITED

The practice of science, outlined above, concurs both with the predominant beliefs
of individual scientists and an epistemology of modest realism. Furthermore, such a
view does have pedagogic implications. The case for realism developed here is based
on the arguments advanced by Harrk (1986) in Varieties of Realism. Harrk sensibly
recognizes that maximal realism, the position of any individual who believes that
theoretical statements are true or false by virtue of the way the world is, is far too
strong and easily vulnerable to attack. Ptolemaic astronomy and continental drift are
just two of many examples which demonstrate that there can be no place in science
for a sure catalog of facts of the matter fixed by the way the world is (Hads

70 OSBORNE
emphasis). Instead, basing his approach on material practice, that is, what scientists
do and say, he recognizes that neither the slogan reference determines sense, or
sense determines reference, which contrast the role of ideas and sensation, will do
as a basis for a theory of meaning. His important contribution is to recognize that:
Sense determines the possibility of reference-that is, it directs a programme of research- while reference, when achieved, brings about refinement of the sense.
There is, over time, a dynamic, dialectic interaction between these two aspects of
meaning. (Harr6, 1986, p. 33)

This position shifts the locus of the argument about realism from the truth or falsity
of statements to the question of the existence or nonexistence of the entities discussed by science. Scientists are not asking Are statements of this theory true or
false? and doing their best to answer but instead ask the question Do things, properties, processes of this sort exist? and do their best within human limitations to find
exemplars. This realist position is epistemically modest and does not make incorrigible existential claims about reality. Rather it bases its knowledge claims on the
fact that it is possible to secure reference to a referent, what Harr6 terms referential
realism. To establish the latter, Harr6 bases his argument on Gibsons theory of perception (Gibson, 1979), which sees the object and the observer as part of an ambient
array or energy flux, shaped by both the presence of the perceiver and the perceived.
This array or flux provides information which is specific to the object, and it is the
structures within the array caused by the object, what Ham6 terms the higher order
invariants, that the organism perceives with its perceptual system. Such sensations
are coarse-grained, but acceptance of a Gibsonian theory of perception explains how
it is possible to perceive something generic-a thing, an event or a spatial relation.
Only then does the cognitive operation enter into the process to interpret and define
the information afforded by the object in the ambient array. If necessary, the organism will use additional sensory organs to establish the nature of the higher order invariants through an interactive process of exploration and feedback.
The important point about the Gibsonian theory of perception is that it avoids the
trap of making representations prior to sensations. For if categories and concepts are
to be determined a priori-from what are they to be constructed? A Gibsonian version of perception shifts the point at which cognitive influences enter the act of perception to after rather than before, a subtle but significant distinction which avoids
the commitment to the transcendental necessity of categories a priori. Thus, a
Gibsonian theory of perception is similarly modest. What is says is that things exist -we can perceive them and they are ontologically stable, but our description and
beliefs about them can be revised.
In arguing for a referential realism, Harr6 states that our interaction with, and sensory feedback from macroscopic objects allows us to know what the attributes of
such entities are. Their ontological status is not in question but the notion that we ultimately can know the truth about such objects is sensibly dismissed as a fruitless
search. All that it attempts to show is that the objects of scientific investigation are
relatively stable as opposed to the cluster of beliefs that we may hold about it.
Consequently, what Lister observes as loose cells in an infected wound and Pasteur

BEYOND CONSTRUCTIVISM 71
describes as external agents of infection exist and can be referred to through material practice. Referential relations and our descriptive vocabulary can be revised but
the basic ontological sketch is not in question. For instance, the heart and the circulatory system existed before Harvey first described them in 1628.
Having established this carefully constructed epistemic position based on successful and unsuccessful acts of reference in which a real physical relation is established between the perceiver and an exemplary thing, HarrCs next significant step
is to argue that there are three types of entities that we experience in the world
which require not a singular theory of science but a triadic one. Realm 1 theories
enable the classification and predictions about macroscopic objects which are tangible and accessible to sensorimotor experiences; thus, a typical realm 1 theory is
classical kinematics-an area of common experience. The moon, the eye, a chair,
a spring are all typical referents belonging to realm 1. Thus, for such objects,
meaning in scientific cognition and discourse is typically provided by reference to
samples which fix the meaning of descriptive terms, and hence is essentially extralinguistic.
Realm 2 theories are iconic in the sense that they represent unobservable entities
which are only accessible to our senses through instrumentation such as bacteria and
viruses. However, such entities and their behaviors are always observable, or potentially so, through the aid and assistance of technical artifacts. All such entities are the
products of a representation of a physical system and its modes of behavior, which
when first formulated, have not been observed. In realm 2 discourse, the objects of
realm 1 serve as a basic ontology, but such discussion is much more dependent on intralinguistic meaning creation through the use of metaphor and simile. Harrk follows
Soskice (1985) in arguing that:
Metaphor is a trope through which new vocabulary is created. By means of it our
conceptual grasp of a subject matter (referent) is enhanced by the use of a term with
a well-established context of use, and thus an existing deep grammar and set of associated commonplaces. . . . In this way, through metaphor, new vocabulary can be
created within the existing structure of language, so securing the intelligibility .of the
term in the context of use. (Harri, 1986, p. 77)

Hence, metaphor is not simply a useful adjunct for scientific thinking but an essential
component of theory itself which leads to the discovery of hidden causes, or the postulation of new theoretical entities. The vast majority of scientific theories are descriptions and hypotheses of realm 2 objects.
Finally, realm 3 theories describe theoretical objects for which there is no direct
evidence of their existence, such as quarks and black holes, whose descriptions are
often mathematical. Such objects are beyond all possible experience but advances in
technology may enable these objects to be observed and thus make an ontological
shift into realm 2. For instance, prior to the invention of the electron microscope,
viruses were unobservable and at the time would have typified the speculative realm
3 entity which the Higgs Boson currently does for particle physicists. Once again,
our knowledge of these objects is dependent on the use of metaphor, but also the language of mathematics.

72 OSBORNE
Hacking (1983) also offers a realist interpretation of science that does make the
important distinction between real and theoretical entities. Entities become real when
they are manipulated by the experimenter who intervenes on the world. Using electrons as an example he argues that because we regularly set out to build, and often
enough succeed in building new kinds of devices that use various well-understood
causal properties of electrons such entities do exist. The essence of his argument is
that the existence of electrons is not a hypothesis tested by the successful construction of some artifact but the reverse-that the effective functioning of the newly
built instrument is dependent on a modest number of home truths about electrons
which are confirmed time and time again. For theories, Hacking advances a pragmatic modest realism following the arguments of Putnam (1975) that theories and
their representations attempt to refer to some reality and science progresses because
these theories improve and our descriptions are enhanced. Although we can never be
certain that they are accurate representations, Hackings realist would argue that the
phrase as certain as we can possibly be summarizes his or her views about the
claims made about reality by theories.
Hacking and H m 6 both present a clear argument for a realist epistemology based
on an objective ontology. This position stands in contrast to those of philosophers
like Rorty (1989) for whom there is nothing to be said about truth save that each of
us will commend as true those beliefs which he or she finds good to believe. Rortys
work i s a serious attempt to dethrone science from a pedestal where it offers a standard of objectivity and rationality to which all other forms of enquiry must be compared. Thus the only way science can be seen as exemplary is that it offers a standard
of practice for the provision of free and open encounter which provides suggestions
about how the rest of culture might organize itself.
As Matthews (1994) argues, modest realism is enough to go on with, and it is incompatible with constructivism. Its limited theses are that there is an ontological reality; that scientific theories are human attempts to refer to this reality; that theories
are confirmable by evidence (or falsifiable); and that there is a progression in scientific theories which produce increasingly accurate approximations and descriptions of
that reality. Such a perspective does have limited pedagogic implications which are
markedly different from those of constructivism.

REALIST PEDAGOGY
HarrCs interpretation of scientific theory accounts for two of the current problems
within science education. First, since much of modem science is concerned with potential realm 3 entities, there is a natural inclination to forget the grounds on which
such objects are formulated-our understanding of realm 1 and 2 objects. The speculative nature of the discussion about black holes or the demise of the dinosaurs leads
some to imagining that all science is simply a way of talking about the world with no
substantive basis for its knowledge claims. The two most well-known attacks on science by the sociologists of science are both studies of scientists working in Cealm
3-in one case on the development of our understanding of quarks (Pickering, 1984)
and the other on the determination of the molecular structure of TRF hormone
(Latour & Woolgar; 1986). Consequently, the view arises that all science is an activ-

BEYOND CONSTRUCTIVISM 73

ity requiring the construction of theoretical entities where the relation to any reality is
at best tenuous or unjustifiable. But this is because the activity of research scientists
working at the frontiers of our understanding with realm 3 entities has been portrayed
as the activity of normative science.
Second, in science education, we have attempted to ontologically shift some entities, such as energy, from realm 3 to realm 1 with a failure to address the inevitable
problems that arise out of attempting to make the intangible concrete and self-evident. The debate about the teaching of energy that has occurred over the past decade
is a good example of the pedagogic problems raised with, on the one hand, the argument that the topic is too difficult and abstract for anybody younger than 18 (Warren,
1983) and, on the other hand, the arguments of others (Carr & Kirkwood, 1988) that
the topic should be included in the primary curriculum.
More importantly, for the science educator, Harrks ideas provide a coherent account of scientific epistemology which is more akin to that of practicing scientist and
some tentative theoretical basis for the construction of a curriculum. Experiences of
realm 1 macroscopic objects are those which the child uses to construct explanatory
schemas of physical and biological phenomena, albeit fragmented and lacking any
theoretical description other than an intuitive mechanics (Mariani & Ogborn, 1991;
Bliss et al., 1989; di Sessa, 1988) and an intuitive biology (Carey, 1985). If so, an
early science education should attempt to build on and extend childrens experiences
of macroscopic phenomena, introducing the child to the descriptive language of the
scientist and the theoretical frameworks, the ideas of realm 2 and 3, which enable
them to generalize from such experiences and to interpret their acts of perception
with the scientists spectacles.
Such an approach would encourage observational activities of a wide range of
common macroscopic phenomena. For instance, the fact that all liquids can be made
to disappear and reappear and that such properties are what scientists call evaporation and condensation; that all animals have a mechanism for taking in oxygen
and exhaling carbon dioxide; that springs and rubber bands stretch and that they all
share a common pattern in the way in which they stretch. Thus, the sole purpose of
such an education at this stage would be to enhance childrens discourse and to show
that, with such a language, all phenomena are not uniquely contextualized but share
universal properties. Moreover, such an approach would place a greater emphasis on
introducing the theoretical description prior to observation and experience rather
than a vain hope that the child might spontaneously discover the scientific explanation from an interpretation of their empirical data.
From here, science education would move to examine the theoretical entities of
science for which there is instrumental evidence-Harrks realm 2 entities.
Typically, this would require an exploration of the particle model and the evidence to
support it; the evidence of microscopy for the internal structure of plants; the evidence of dissection or models for the internal mechanisms of the body; the evidence
for charged nature of matter; the nature of the Solar System, etc. All are entities
whose ontological existence is inferred from the evidence of instrumentation.
However, their existence can be related to realm 1 macroscopic objects, the entities
of possible experience for children, through the use of direct reference and metaphor
to construct models and representations.

74 OSBORNE

Realm 3 entities are the abstractions of the human mind for which there is no instrumental evidence. As such these entities are inferred to explain certain experimental results-neutrinos and quarks are both classic examples. However, at a more
mundane level, quantities such as speed, acceleration, current, charge, energy, the
mole, and molecular biological mechanisms are such abstractions. These concepts
can only be constructed on a sound scientific understanding of realm 1 and 2 entities.
Hence, the idea of speed is constructed out of the pupils experiences of motion and
the measurement of time and distance, both realm 1 entities and the idea of current
constructed from observations of the brightness of bulbs and simple causal relationships. Cognitively such ideas are more demanding because they require the ability to
envisage and manipulate imagined entities and their symbols for which there is no direct referent.
In this way, HarrCs account of science provides a form and logic for the curriculum that bears a clear and transparent relationship to the structure and practice of science itself. An epistemology of modest realism insists that knowledge claims are
checked against reality itself and thus avoids the relativist pitfall of the radical constructivist who would ascribe validity to any viable theory that encapsulated personal experience. More importantly, it would enable the curriculum developer to
make decisions about content, a matter on which a constructivist epistemology says
nothing. Yet, quite plainly, we choose not to teach children about special relativity,
quantum mechanics, molecular biology, and the electronic configuration of chemical
bonds because we know that such knowledge is not accessible to them until they
have understood a wide body of other factual and theoretical information.
Constructivism singularly fails to provide any reason why this should, or should not,
be done.
STYLES OF LEARNING
Finally, there is one other issue that merits consideration which is the nature of the
individual reaction to the style of learning offered by constructivist pedagogy. In reacting against didacticism, constructivists in science education have chosen to place
their emphasis on cooperative, discussion-based activities for the production of
knowledge and ignore the fact that there may be a role for demonstrating, showing,
or telling. For instance, Tobin (1990) argues for a learning model based on cooperative learning and social collaboration; the curriculum development work undertaken
by Driver (1989) uses small groups of about four students which form the structural unit around the scheme of activities; and Nussbaum and Novak (1981) advocate the use of brainstorming in the classroom as a pedagogic strategy. Others
place an emphasis on interpretive discussion or listening discussion (Baird &
Northfield, 1992) and group activities such as games, role playing, and simulations
(Bentley & Watts, 1989). But the evidence of research would indicate that the approach to learning offered by a constructivist pedagogy is preferred by some, but not
all students, and likewise, is effective for some but not all.
The literature on learning styles is extensive. Brophy and Good (1974) explored
the effects of diverse cognitive styles and personality characteristics, while Good and
Power ( 1976) examined the effect of affective characteristics (e.g., attitudes, interest,

BEYOND CONSTRUCTIVISM 75
and motivation). However, some of the most interesting work has been undertaken by
Pask (1976) and Entwistle (1981). Pask identified two main types of learning strategy
used by different individuals: the serialist, who deployed a step-by-step strategy
which examined one hypothesis and then the next in a simple linear progression; and,
in contrast, the holist, who took a more global approach to problem solving, considering multiple hypotheses simultaneously, and used a more individualistic approach to learning. Pask found that both groups of students were capable of reaching
the same level of understanding, but that their ways of attaining that understanding
were very difSeerent. Holists prefer to start by forming an overall picture of what is being learned while serialists attempt to integrate separate topics in a piecemeal form.
Hence, while the holist is concerned with comprehension learning, constructing descriptions of what is known, the serialist takes an operational approach, attempting
to master procedural details.
One of the most important experiments conducted by Pask, which those who
wholeheartedly advocate a constructivist pedagogy should note, was to match and
mismatch sets of learning materials with students learning styles. The students in the
matched conditions were able to answer most of the questions about what they had
learned, whereas the mismatched students generally achieved below half marks.
Entwistles work explored the personality factors that correlated with academic
success in undergraduates. He identified three distinct personality types, whose approaches to learning were all significantly different and who achieved success. The
first group consisted of motivated students who were emotionally stable and were
spurred by competition to demonstrations of intellectual mastery. The second group
were the antithesis in that they were unrealistically pessimistic about their ability and
haunted by the fear of failure. Driven by anxiety, these students had unconventional,
though effective, approaches to studying. The final group were predominantly artsbased students and combined high verbal aptitude with good study methods and long
hours of study.
The important conclusion to be drawn from this kind of research is that students
vary in their motivation and preferred learning styles, and that a teaching scheme
based on a single perspective will only meet the needs of a subset of any group of
children. Within science education itself, there are studies which support such an interpretation.
For instance, in a study of 390 Spanish 15-year-olds learning styles, Kempa and
Martin-Diaz (1990a, 1990b) identified four types of motivational patterns, driven out
of a desire to either: (a) achieve; (b) satisfy their curiosity; (c) fulfill or discharge a
duty; or (d) affiliate and interact with other people. These they call respectively the
achievers, the curious, the conscientious, and the social. Using an 80-item
Likert-type preference inventory, they then examined the relationships that existed
between these motivational patterns and their liking for different instructional procedures. Only the social learners showed a strong preference for group activity and
the opportunity to pursue ones own inquiry. In contrast, conscientious learners actually exhibited a preference for didactic teaching and experimental work with instructions. Curious learners preferred learning from books, discovery learning, and
doing practical work. Achievers had no strong preferences for any mode of learning. The clear point is that individuals vary in their preferred learning style and no

76 OSBORNE

one model will meet the diverse needs of different children. Further evidence that
this is so can be found in the report of the PEEL Project where a pupil comments:
Its about us learning about, and comprehending the work that teachers set. . . .
helping students learn better by revising the days work, and writing down in your
diary what you have done. , . . I didnt like doing it-I stopped about the second
week we did it-it wasnt interesting. (Baird & Northfield, 1992, p. 57)

That not all pupils find such an approach to their liking is important as this implies
that there is no one strategy that will achieve success with all pupils. Kempa and
Martin-Diaz argue that the only practicable solution to this problem is for teachers
to use as wide a range [emphasis added] of instructional procedures as possible, instead of limiting themselves to one or two. Yet the danger in constructivist pedagogy is an assumption that it offers an improved learning strategy for all pupils,
whereas its major strength lies in offering an alternative-challenging teachers wedded to a didactic model of transmission by offering variety and diversity.
Therefore, the evidence from research on learning styles would suggest that there
is no single, effective method for teaching and learning as students differ in their preferences. From a psychological perspective and a careful examination of the aims of
teaching science, Claxton ( 1993) argues convincingly that a single-minded emphasis
on conceptual development in science is inappropriate for all children. Instead, science education should be developing a wide repertoire of teaching methods that are
custom-built for different aims and different clientele. Consequently, the only tenable position to hold in science education is one which sees it as an organic and indeterminate process where the epistemic biography of each individual is unique and
also uniquely determined.

CONCLUSIONS

No doubt there will be some who will dispute the many points made in this critique. Some will find the notion of a realist epistemology per se difficult to accept,
even with its modest limitations. For others its clear relationship to the structuralism
of genetic epistemology will engender doubt in many minds. However, the argument
has been advanced here that constructivist alternative offers a flawed epistemology
which misrepresents science as it is practiced. There can be no argument with its thesis that all knowledge is a human construction. In that sense we are all trivial conIn
structivists (Glaserfeld, 1993). But that thesis is what it claims-trivial.
advancing a more elaborated epistemological argument, it has made the mistake of
focusing on the notion of what scientists working at the leading edge engaged in the
construction of new theoretical entities do, rather than looking at the work of scientists and engineers engaged in normal science who repetitively and routinely confirm
that these ideas have predictive validity and add to our knowledge of objects that initially seem tenuous. This has lead to the portrayal of science as a process of constructing and manipulating representations which bear no necessary relation to any
ontological reality. In so doing, constructivists have forgotten that it is the world that
imposes constraints on human thought, and not human thought that imposes con-

BEYOND CONSTRUCTIVISM 77
straints on the world. Furthermore, constructivisms focus on the social and human
construction of knowledge has resulted in an instrumental view of theories, which requires only that they be viable, and a failure to consider and elaborate any criteria
for the adjudication of competing knowledge claims. In contrast, the case has been
made that the modest realist position elaborated by Harr6 (1986) offers a view of science which is more commensurate with that of its practitioners and, in addition, offers a limited set of implications for the content and sequencing of the science
curriculum.
Basically there are two sources of human learning-knowledge that is acquired
through sensorimotor interaction, that is, by acting and intervening on the world, and
knowledge that this is acquired through cultural transmission, be it through the popular media or specialized institutions such as schools. Constructivist research has been
seminal in exploring the learning outcomes resulting from the first category, and in
beginning to explore the nature of new learning experiences that would enable the
child to understand the standard scientific worldview. However, in its reaction against
didacticism, it has ignored the important issue of how the ideas of science may be
told or shown to children, and instead offered a singular pedagogy where meaning is
negotiated through a process of cultural apprenticeship. For constructivism would appear to hold the belief that an objective view of knowledge inevitably requires a didactic pedagogy and by inference, to advance any role for transmitting knowledge
would be to present it as objective. There is no logical justification to this premise
and the two are not inextricably linked.
Constructivism has been invaluable in attempting to untie the Gordian knot between objectivism and didacticism but it is just that-a knot to be untied, based
on habit and past practice, rather than a permanent or logical connection.
Similarly the realist perspective can, and should, offer a pluralist pedagogy as
should constructivism. For, in the absence of telling or showing, from whence are
students to gain the scientists understanding? Sciences theoretical constructs are
critical in enabling common experiences to be reinterpreted. The notions that the
earth spins, that inherited characteristics are transmitted through coded information on genes, and that all matter is made of a limited number of elements, are not
self-evident. Science is fundamentally unnatural (Wolpert, 1992) and not common-sensical (Cromer, 1993). Therefore, children must be shown and introduced
to ideas that are not palpable and to the advantages that such concepts bring in understanding our world.
A typical example of the problem can be found in the long list of learning activities in the UK Association for Science Education Science Teachers Handbook
(Ramsden & Harrison, 1993). The list contains such elements as raising questions,
making observations, using practical skills, and small group discussions, yet not one
mention is made of an activity which would enable students to be provided with a
scientific theory. Such authors would seem to have forgotten Hodsons (1990) point
that:

. . . the simple matter is that theoretically uninformed observations do not

[Hodsons emphasis] and cannot lead to the acquisition of new concepts. The claims
for theory-free experimentation are nonsensical on both epistemological and psycho-

78 OSBORNE
logical grounds. . . . In short, theoretical considerations must precede experimental
enquiry. (Hodson, 1990, p 37)

The uncritical adoption of constructivism as a referent, on the basis of its achievements as a method, has led to the evolution of arguments which attempt to deny the
value of some didactic processes to some students, an epistemology which has no
statement to make about developmental sequences or curriculum content and a pedagogy which lacks an adequate theoretical underpinning. While there is much value to
this school of thought and the products of its research, these and other criticisms have
to be made because:

. . . no one metaphor should be allowed to prevail (as currently the information

processing metaphor prevails in cognitive psychology) and to become discursively
hegemonic. . . . New lenses of many different kinds would now seem to be
needed-even if we can no longer supply the single Gods eye view of things
desired-the craft of the intellectual lens-maker to the public at large would still
seem to hold some honour. (Shotter, 1992, p. 11)
Therefore, an improved science education will only come through the critical review
of arguments and research evidence, by the adoption of a pedagogy which places a
value on variety and diversity, and not through adherence to a singular school of
thought.
I am indebted to Martin Monk, Kings College London, for his comments on several drafts
of this article and for first introducing me to Harr6s realist epistemology, and to Justin Dillon
and Paul Black for their help and comments.

REFERENCES
Adey, P., & Shayer, M. (1994). Really raising standards. London: Routledge.
Ausubel, D. P. (1968). Educational psychology: A cognitive view. New York: Holt, Rinehart &
Winston.
Baird, J. R., & Northfield, J. R. (Ed.) (1992). Learningfrom the Peel experience. Melbourne:
University of Monash.
Bentley, D., & Watts, M. (Ed.) (1989). Learning and teaching in school science. Milton
Keynes, UK: Open University Press.
Bliss, J., Ogbom, J., & Whitelock, D. (1989). Secondary pupils commonsense theories of motion. International Journal of Science Education, l l , 261 -272.
Brophy, J. C., & Good, T. L. (1974). Teacher-student relationships: Causes and consequences. New York: Holt, Rinehart & Winston.
Brown, A. (1987). Metacognition, executive control, self-regulation, and other more mysterious mechanisms. In F. Weinert & R. H. Kluwe (Eds.), Metacognition, motivation and understanding. Hillsdale, NJ: Lawrence Erlbaum, pp. 65- 113.
Brown, J., Collins, A., & Duiguid, P. (1989). Situated cognition and the culture of learning.
Educational Researcher, 18,32-42.
CLIS Project (1987). CLIS in the classroom. Leeds: Centre for Science and Mathematics
Education, University of Leeds.
Carey, S. (1985). Conceptual change in childhood. Cambridge, MA: MIT Press.

BEYOND CONSTRUCTIVISM 79
Carr, M., & Kirkwood, V. (1988). Teaching and learning about energy in New Zealand secondary school junior science classrooms. Physics Education, 22,86-91.
Claxton, G. (1993). Interplay of values and research. In P. J. Black & A. M. Lucas (Eds.),
Childrens informal ideas about science. London: Routledge, pp. 190-207.
Collins, H. M. (1985) Changing order: Replication and induction in scientific practice.
London: Sage.
Cosgrove, M., Osbome, R., & Can; M. (1984). Childrens intuitive ideas on electric current
and the modification of those ideas. In R. Duit, W. Jung, & C. Rhoneck (Eds.), Aspects of
understanding electricity. Ludwigsburg : Institut fur die padagogik der naturwissenschanften
and der Universitat Kiel, pp. 247-266.
Cromer, A. (1993). Uncommon sense: The heretical nature of science. New York: Oxford
University Press.
Dana, T. M., & Davis, N. T. (1993). On considering constructivism for improving mathematics and science teaching and learning. In K. Tobin (Ed.), The practice of constructivism in
science education. Washington, DC: AAAS Press, pp. 220-234.
Davies, F., & Greene, T. (1984). Reading for learning in the sciences. Edinburgh: Oliver &
Boyd.
di Sessa, A. A. (1988). Knowledge in pieces. In G. Forman & P. B. Pufall (Eds.),
Constructivism in the computer age. Hillsdale, NJ: Lawrence Erlbaum.
Driver, R. (1989). Changing conceptions. In P. Adey, J. Bliss, J. Head, & M. Shayer (Eds.),
Adolescent development and school science. London: Falmer Press, pp. 79-99.
Driver, R., & Easley, J. (1978). Pupils and paradigms: A review of literature related to concept
development in adolescent science students. Studies in Science Education, 5,61-84.
Driver, R., & Oldham, V. (1985). A constructivist approach to curriculum development.
Studies in Science Education, 13, 105- 122.
Driver, R., Asoko, H., Leach, J., Mortimer, E., & Scott, P. (1994a). Constructing scientific
knowledge in the classroom. Educational Researcher, 23,5 - 12.
Driver, R., Squires, A., Rushworth, P., & Wood-Robinson, V. (1994b). Making sense of secondary science. London: Routledge.
Duit, R. (1993). Research on students conceptions-developments and trends. Paper presented at the Third International Seminar on Misconceptions in Science and Mathematics
Education. Cornell University, Ithaca, NY.
Duschl, R. A. (1990). Restructuring science education. New York: Teachers College Press.
Entwistle, N. (1981). Styles of learning and teaching. Chichester: Wiley.
Gibson, J. J. (1979). An ecological approach to visual perception. London: Houghton Mifflin.
Giere, R. (1984). Understanding scientific reasoning (2nd ed.). New York: Holt, Rinehart &
Winston.
Gilbert, J. K., Osborne, R. J., & Fensham, P. J. (1982). Childrens science and its consequence
for teaching. Science Education, 66,623 -633.
Glaserfeld, E. V. (1987). Learning as a constructive activity. In C. Janvier (Ed.), Problems of
representation in the teaching of mathematics. Hillsdale, NJ: Lawrence Erlbaum, pp. 3- 17.
Glaserfeld, E. V. (Ed.) (1989). Constructivism in education. Oxford: Pergamon Press.
Glaserfeld, E. V. (1991). Introduction. In E. V. Glaserfeld (Ed.), Radical constructivism in
mathematics education. Dordrecht: Kluwer, pp. xxiii- xx.
Glaserfeld, E. V. (1993). Questions and answers about radical constructivism. In K. Tobin (Eds.),
The practice of constructivism in science education. Washington, DC: AAAS Press, pp. 23-38.
Good, T. L., & Power, C. N. (1976). Designing successful classroom environments for different types of student. European Journal of Science Education, 7,221 -229.
Gunstone, R. F. (1991). Reconstructing theory from practical experience. In B. Woolnough
(Ed.), Practical science. Buckingham, UK: Open University Press.

80 OSBORNE
Gunstone, R. F., & White, R. T. (1981). Understanding of gravity. Science Education, 65,
29 1-299.
Hacking, I. (1 983). Representing and intervening. Cambridge: Cambridge University Press.
Harr6, R. (1986). Varieties of realism: A rationale for the natural sciences. Oxford: Basil
Blackwell.
Harris, D., & Taylor, M. (1983). Discovery learning in school science: The myth and the reality. Journal of Curriculum Studies, 15,277-289.
Hewson, P. W. (1981). A conceptual change approach to learning science. European Journal
of Science Education, 3,383-396.
Hewson, P. W., & Hewson, M. G. A. (1984). The role of conceptual conflict in conceptual
change and the design of science instruction. Instructional Science, 13, 1- 13.
Hodson, D. (1988). Towards a philosophically more valid science curriculum. Science
Education, 72, 19-40.
Hodson, D. (1 990). A critical look at practical work in school science. School Science Review,
70,33-40.
Horton, R. (1971). African traditional thought and Westem science. In M. D. Young (Ed.),
Knowledge and control. London: Collier Macmillan.
Horton, P. B. (1992). An investigation of the effectiveness of concept mapping as an instructional tool. Science Education, 77,95- 111.
Inglehart, R. (1990). Culture shift in advanced industrial society. Princeton, NJ: Princeton
University Press.
Kelly, G. A. (1955). The psychology of personal constructs. New York: W. W. Norton.
Kempa, R. F., & Martin-Diaz, M. (1990a). Motivational traits and preferences for different instructional modes in science. Part 1: Student motivational traits. International Journal of
Science Education, 12, 194-203.
Kempa, R. F., & Martin-Diaz, M. (1990b). Students motivational traits and preference for different instructional modes in science. Part 2. International Journal of Science Education,
12,205-216.
Koulaidis, V., & Ogbom, J. (1995). Science teachers philosophical assumptions: How well do
we understand them? International Journal of Science Education, 17, 273-282.
Latour, B. (1985). Science in action: How to follow scientists and engineers through society.
Cambridge, MA: Harvard University Press.
Latour, B., & Woolgar, S. (1986). Laboratory life: The construction of scientific facts (2nd
ed.). Princeton, NJ: Princeton University Press.
Laudan, L., Donovan, A., Laudan, R., Barker, P., Brown, H., Leplin, J., Thagard, P., &
Wykstra, S. ( 1986). Scientific change: Philosophical models and historical research.
Synthese, 69, 141-223.
Lave, J. (1988). Cognition in practice: Mind, mathematics and culture in everyday life.
Cambridge: Cambridge University Press.
Lemke, J. L. (1990). Talking science: Language, learning and values. Norwood, NJ:
Ablex.
Lyotard, J. F. (1984). The postmodern condition: A report on knowledge. Minneapolis, MN:
University of Minnesota Press.
Margetson, D. (1982). Science education and the philosophy of science. Research in Science
Education, 11, 1-7.
Mariani, M. C., & Ogborn, J. (1991). Towards an ontology of common-sense reasoning.
International Journal of Science Education, 13,69-85.
Matthews, M. (1995). Constructivism and New Zealand science education. Auckland:
Dunmore Press.

BEYOND CONSTRUCTIVISM 81
Matthews, M. R. (1994). Science teaching: The role of history andphilosophy of science. New
York: Routledge.
Millar, R. (1995, April). Knowledge and action: Students understanding of the procedures of
scientific inquiry. Paper presented at the First European Conference for Research in Science
Education, Leeds.
Muthukrishna, N., Carnine, D., Grossen, B., & Miller, S. (1993). Childrens alternative frameworks: Should they be directly addressed by science instruction: Journal of Research in
Science Teaching,30,233-248.
Nola, R. (1995a, in press). There are more things in heaven and earth, Horatio than are dreamt
of in your Philosophy: A dialogue on realism and constructivism. Studies in the History and
Philosophy of Science.
Nola, R. (1995b). Against constructivism in the philosophy of science education. Private communication.
Novak, J. D., & Gowin, D. B. (1984). Learning how to learn. Cambridge: Cambridge
University Press.
Nussbaum, J., & Novak, J. (1981). Brainstorming in the classroom to invent a model: A case
study. School Science Review, 62, 771 -778.
Ogbom, J. (1988). A map of science. Personal submission to the National Curriculum Working
Group on Science Education.
Ogbom, J. (1994) Theoretical and empirical investigations of the nature of scientific and commonsense knowledge. Unpublished doctoral thesis, Kings College, London.
Osbome, J. F. (1990). Sacred cows in physics-towards a redefinition of physics education.
Physics Education, 25, 189- 195.
Osbome, R. J., & Wittrock, M. (1985). The generative learning Model and its implications for
science education. Studies in Science Education, 12,59-87.
Pask, G. (1976). Styles and strategies of learning. British Journal of Educational Psychology,
46, 128-148.
Pickering, A. (1984). Constructing quarks. Edinburgh: Edinburgh University Press.
Pope, M. (1985). Constructivist goggles: Implications for process in teaching and learning.
Paper presented at the annual conference of the British Educational Research
Association.
Putnam, H. (1975). Mind, language and reality. Cambridge: Cambridge University Press.
Ramsden, P., & Harrison, B. (1993). Teaching science. In R. Hull (Eds.), Association for
Science Education Science Teachers Handbook: Secondary. Heme1 Hempstead, UK:
Simon & Schuster.
Rorty, R. (1989). Contingency, irony and solidarity. Cambridge: Cambridge University
Press.
Rosen, E. (1962). Three Copernican treatises. New York: Dover.
Rowell, J. A., & Dawson, C. J. (1983). Laboratory counterexamples and the growth of understanding in science. European Journal of Science Education, 5,203 -215.
Sainsbury, M. ( 1992). Meaning, communication and understanding in the classroom.
Aldershot: Avebury.
Samarapungavan, A. ( 1992) Childrens judgements in theory choice tasks: Scientific rationality in childhood. Cognition, 45, 1-32.
Shavelson, R. J. (1974). Methods of examining representations of a subject-matter structure in
students memory. Journal of Research in Science Teaching, I I , 2 3 1-249.
Shayer, M., & Adey, P. (1981). Towards a science of science teaching. London: Heinemann.
Shotter, J. (1 992). In dialogue: Social constructionism and radical constructivism. Paper presented at the conference on Alternative Epistemologies in Education, Athens, GA.

82 OSBORNE
Shulman, L. S. (1986). Those who understand: Knowledge growth in teaching. Educational
Researcher, 15,4- 14.
Solomon, J. (1994). The rise and fall of constructivism. Studies in Science Education, 23,
1 - 19.
Soskice, J. H. (1985). Metaphor and religious language. Oxford: Clarendon Press.
Stavy, R., & Berkovitz, B. (1980). Cognitive conflict as a basis for teaching quantitative aspects of the concept of temperature. Science Education, 64,679-692.
Stoppard, T. (1993).Arcadia. London: Faber and Faber.
Thorley, N. R., & Treagust, D. F. (1986). Conflict within dyadic interactions as a stimulant for
conceptual change in physics. International Journal of Science Education. 9, 2, 203-216.
Tobin, K., & Tippins, D. (1993). Constructivism as a referent for teaching and learning. In K.
Tobin (Ed.), The practice of constructivism in science education. Washington, DC: AAAS
Press, pp. 3-21.
Tobin, K., Butler Kahle, J., & Fraser, B. J. (1990). Learning science with understanding: In
search of the Holy Grail? In K. Tobin, J. Butler Kahle, & B. J. Fraser (Eds.), Windows into
science classrooms London: Falmer Press.
Tobin, K., Tippins, D. J., & Gallard, A. J. (1994). Research on instructional strategies for
teaching science. In D. L. Gabel (Ed.), Handbook of research on science teaching and
learning. New York: Macmillan, pp. 45-93.
Warren, J. W. (1983). Energy & its carriers: A critical analysis. Physics Education, 18,
209 -2 12.
Weinberg, A. M. (1968). The two faces of science. Journal of Chemical Education, 45,74.
White, R. T. (1988).Learning science. Oxford: Basil Blackwell.
White, R., & Gunstone, R. (1992).Probing understanding. London: Falmer Press.
White, R. T., & Mitchell, I. J. (1994). Metacognition and the quality of learning. Studies in
Science Education, 23,21-37.
Wolpert, L. (1992). The unnatural nature of science. London: Faber and Faber.
Woolnough, B., & Allsop, T. (1986). Practical work in science. Cambridge: Cambridge
University Press.
Ziman, J. M. (1979).Reliable knowledge. Cambridge: Cambridge University Press.

Accepted for publication 25 July 1995