Content uploaded by Anna-Mari Wallenberg
Author content
All content in this area was uploaded by Anna-Mari Wallenberg on Aug 20, 2014
Content may be subject to copyright.
Conceptual Coherence in Philosophy Education - Visualizing Initial Conceptions
of Philosophy Students with Self-Organizing Maps
Anna-Mari Rusanen (anna-mari.rusanen@helsinki.fi)
Philosophy of Science Group, Department of Philosophy, PO BOX 9
00014 University of Helsinki, FINLAND
Otto Lappi (otto.lappi@helsinki.fi)
Cognitive Science Unit, Department of Psychology, PO BOX 9
00014 University of Helsinki, FINLAND
Timo Honkela (timo.honkela@tkk.fi)
Adaptive Informatics Research Center, PO BOX 5400
FI-02015 TKK Helsinki University of Technology, FINLAND
Mikael Nederström (mikael.nederstrom@psycon.fi)
Department of Psychology, PO BOX 9
00014 University of Helsinki, FINLAND
Abstract
We present a framework for research on coherence of student
conceptions in philosophy education. Commonsense
conceptions of philosophical novices were studied. Students
of a Finnish upper secondary school with no prior background
in philosophy were asked to evaluate statements on
conceptual issues in the domains of philosophy of mind,
metaphysics and epistemology. The results were visualized
with Kohonen self-organizing- maps (SOM), enabling us to
identify clusters of students and questions with similar
response patterns. The results are interpreted in terms of
students’ ontological commitments.
Keywords: Conceptual change; commonsense conceptions;
philosophy of mind; self-organizing maps.
Introduction
Most students have an intrinsic interest in philosophical
problems. They spontaneously develop their own
conceptions, arguments and theories about various
philosophical issues, including metaphysical and ontological
questions about the nature of mind. What is known of these
students’ ideas of philosophical problems? What are the
crucial cognitive and metacognitive operations required to
comprehend philosophical content? What are the student
characteristics enabling them to - or hindering them from –
developing coherent philosophical viewpoints and sustained
arguments on philosophical topics?
This paper has two main foci. We outline a framework
where the empirical research of conceptual organization of
philosophy novices could be placed, and we also present the
results of our own preliminary empirical research along
these lines.
The general theoretical approach of this study is based on
the conceptual change paradigm. According to it the starting
point of the learning process are the commonsense theories
people have developed prior to instruction, and the outcome
of successful learning is internalization of a philosophical
theory or a worked-out philosophical position on the issues.
There is a large body of research which shows that novi-
ces conceptions do differ from those of experts’, but re-
searchers still remain divided not only about the nature of
those differences, and also the status of novices’ belief sys-
tems. Some researchers claim that novices belief systems
are weakly organized systems that are internally inconsis-
tent, piecemeal and incoherent (diSessa, 1993). Other re-
searchers argue that novice belief systems are not only
internally quite coherent but they may also share the
essential properties of scientific theories (Chi, 1992;
Samarapungavan and Wiers 1997; Vosniadou & Brewer,
1992).
In our empirical design we focused on the initial state of
the learning process – the commonsense conceptions of
philosophical topics students hold when they are first
exposed to academic philosophy1. The aim of the study was
to capture some features of philosophy novices’ conceptual
organization, and also to give us clues of the degree of their
belief systems’ coherence. In this paper, we will first take a
brief look at the theories of conceptual change (section two)
and then present our empirical research (section three).
Background: Conceptual Change and
Philosophy, some Domain specific
considerations
In the conceptual change paradigm, learning of scientific
content is seen as a replacement of confusions and mis-
conceptions or everyday commonsensical frameworks with
new more sophisticated and theoretically deeper ones. This
is taken to entail some kind of wholesale reconstruction of
1 In Finland this occurs at the first year of upper secondary
school.
1674
one's theoretical outlook on a domain. In conceptual change
the difference between the initial state and the outcome of
learning is not merely accumulation of knowledge and
rejection of false beliefs. Instead, the students’ conceptions
of phenomena in a domain undergo a holistic restructuring
process, leading to acquisition of new (for the students)
scientific concepts and a reorganization of the students’ web
of beliefs from a fragmented set of commonsense beliefs to
a consistent web of scientific conceptions. (Carey, 1985;
Strike & Posner, 1982; Vosniadou, 1992; Vosniadou &
Brewer, 1992; diSessa, 1988, 1993; Chi, 1992).
In general, conceptual change is considered as a replace-
ment of old, naïve conceptions with the new, more sophisti-
cated ones. This replacement process is usually thought to
require some sort of cognitive conflict, since there must be
dissatisfaction with the existing concepts (Strike and Posner,
1982). Moreover, the process requires also an alternative
conception that is intelligible, appears initially plausible and
is believed to be a fruitful way to conceptualize the domain
(Strike and Posner, 1982). In science education the domain
of interest could be Newtonian motion or biological
inheritance and evolution. The outcome would be then an
understanding of the relevant scientific theories.
The assumption is that by confronting the belief system
with real world counterexamples, the misconceptions - i.e.
sources of confusion and incoherence in the commonsense
picture - can be revealed. This kind of procedure reflects not
only the basic methodological principles but also the
explanatory aims of these experimental sciences. These
sciences aim to describe the relevant causal or functional
organization in the world (see Cummins, 1983, Bechtel &
Richardson, 1993, Craver, 2006). Their theories and models
are taken to represent the world in a sufficiently correct
way. Moreover, these explanatory structures are also
intended to help us to understand how the target system
would behave under a variety of causal manipulations
(Woodward, 2003).
In philosophy education the relevant domains and the
outcomes of learning are somewhat harder to delineate. But
it seems natural enough to consider traditional philosophical
topics such as the nature of mind as domains, and well
thought-out philosophical positions as the potential out-
comes of learning. However, philosophy also uses its own
method of inquiry - conceptual analysis rather that empirical
investigation or formal proof – and as the method of theory
formation differs from the method of theory formation used
in the natural sciences. These special characteristics of
philosophy should be taken into account when the learning
of philosophical content matter is analyzed.
In natural science domains cognitive conflicts can be
raised by contrasting and confronting the belief system with
the counterexamples based on empirical data and observable
phenomena in the real world. In contrast, philosophy aims
to uncover tacit commitments, make precise ambiguous
positions, and clarify the interrelationships between
concepts and beliefs. In many cases, philosophical theories
are not about the independently observable structure of the
world at all. Consequently, in philosophy cognitive conflicts
cannot be raised just by confronting the student’s
conceptions with observation or “the world”. A parallel
observation has been be made in the context of other
educational domains, such as history. For instance, Limòn
and Carraterro (1999) have pointed out that the domain of
history allows methodologically an interpretation of
historical events from different perspectives. Hence there is
not one, correct empirical fact that could serve as a source
of correct explanation.
In philosophy misconceptions are more or less purely
conceptual confusions, inconsistent definitions, or logical
fallacies. Philosophical examples and counterexamples are
usually thought experiments interpreted, and then weighed
against intuitions - the students own intuitions and the
intuitions of philosophers putting forward and arguing for a
specific theory.
Conceptual Organization and Coherence
There is a large body of experimentally based literature
where it has been argued that the difference between the
consistency or coherence of the belief systems of novices
and experts is one of degree, not of kind. It has been
proposed, for instance, that the novices “explanatory
frameworks” are internally coherent, consistent, interrelated
set of beliefs that are similar to kuhnian paradigms
(Samarapungavan & Wiers, 1997). Also some researchers,
such as Chi and her colleagues, argue that the belief systems
of novices demonstrate conceptual coherence at the
ontological level (Chi, 1992; Chi & Slotta, 1993; Slotta et
al., 1995). According to Chi people assign entities to
ontological categories, and depending on the category
certain attributes may or may not be attributed to the
objects. For instance, physical objects are assigned to the
category of “matter” and assigned such attributes as
“occupies space”. Chi and her colleagues propose that this
categorization system may offer a coherent basis for the
conceptual organization of novices. In their account,
ontological categories and the related attributes provide
conceptual unity, and pre-theoretical coherence within
domains, across situations.
However, there are researchers, who disagree with this
“theory-theory” characterization of novices’ belief systems.
For instance diSessa (1993) describes novice knowledge as
a weakly organized system of beliefs that is highly context
dependent and internally inconsistent, thereby lacking
internal coherence. In diSessa´s account commonsense
physical knowledge is organized into p-prims, empirical
typologies or low-level abstractions of everyday experience.
Opposing this, Chi and others (Chi & Slotta, 1993; Chi,
Slotta and deLeeuw, 1994) argue that novices responses to
physical problems do in fact demonstrate conceptual
coherence at the ontological level, even in cases where
reasoning is guided by p-prims. Chi and Slotta (1993)
suggest that ontological attributes - such as “occupies
space” - may manifest themselves, in a coherent way,
1675
withiin p-prims - such as “blocking” - which are context
specific descriptions of phenomenological situations.
All boils down to the notion of “coherence”. Conceptual
coherence is the “glue” that holds concepts together in a
web of belief (Murphy and Medin, 1985). There is, of
course, a long tradition in cognitive science research on
coherence. For instance, conceptual coherence can be
defined as follows (from Thagard & Verbeugt, 1998,
Thagard et al, 2002): (i) Conceptual coherence is a
symmetric relation between the pairs of concepts, (ii) a
concept coheres with another concept if they are positively
associated i.e. if there are objects to which they both apply,
(iii) the applicability of a concept to an object may be given
perceptually or by some other reliable source, (iv) a concept
incoheres with another concept, if they are negatively
associated, i.e. if an object falling under one concept tends
not to fall under the other concept. Finally (v) the
applicability of a concept to an object depends on the
applicability of other concepts.
Now, the empirical issue is, to what extent novice
conceptions fulfill these conditions. For instance, if a novice
uses the notions of “matter” and “non-material” in a
coherent way, a novice should be able to formulate a correct
positive association between concepts in cases where there
are objects to which “material” and “non-material” both
apply (i-iii) and also be able to distinct the cases when an
object falling under one concept (“material”) do not fall
under the other concept (“non-material) (iv). Now, the
condition v implies that the applicability of a concept, such
as “material” depends on the applicability of other concepts,
such as “occupy space” etc. The challenge is to
operationalize the notion of coherence in an empirical
design. We present work in this direction.
Methods
To begin an investigation into students’ pre-instruction
ontological categorizations and conceptual coherence in
metaphysics and ontology, we administered a multiple
choice questionnaire on these topics to students of a large
Finnish Upper secondary school for adults. The
questionnaires were designed to include two sets of
questions: one to probe the ontological commitments of the
respondent, and one where the respondents were asked to
suspend their ontological commitments and reason
hypothetically. (The idea with the latter questions was to
force the students to respond on the basis of their intuitive
conceptual commitments rather than ontological beliefs per
se). The students’ responses were coded in the binary format
and fed into a Kohonen Self-Organizing Map (SOM). The
SOM is a neural network that uses an unsupervised learning
algorithm that can be used to visualize hidden statistical
structure in the data, enabling us to demonstrate similarity
groupings of questions and students for further analysis.
Subjects
The subjects (n=68, 38 male 30 female) were students of a
large upper secondary school in the area of Helsinki. They
had no previous background in academic philosophy. The
age of the subjects ranged from 15 to 23 years (mean 18
years, s.d. ±2 years, median 18 years). All were enrolled in
the first, compulsory philosophy course in the Finnish upper
secondary curriculum, and had completed the Finnish
comprehensive school. The students were of variable
socioeconomic background and academic ability.
Materials
To obtain information on the students’ conceptions, we
developed a multiple choice-questionnaire. The paper and
pencil- questionnaire included 63 thematically selected
items. This paper presents the results of three thematic sets
of questions. These sets probe (1) the subjects’ ontological
commitments with regard to the mind and the body, (2)
hypothetical questions that relate to the possible spatial and
temporal attributes of bodyless minds (3) hypothetical
questions that relate to the possible perceptual and cognitive
attributes of bodyless minds. Each of these conceptual
subdomains was probed with multiple questions, and the
students’ responses were examined, coded in the binary
format and used to train a self- organizing map for
visualization.
Data-analysis
The self-organizing map is an artificial neural network that
uses an unsupervised learning algorithm as the basis of
similarity comparisons. The SOM, in a nutshell, analyzes
the hidden structure of the input data. More precisely, the
architecture of a Kohonen self-organizing map can be
summarized as follows (from Kohonen, 2001). It consists of
a set of interacting adaptive processing elements, adaptive
prototypes. They are usually arranged as a two-dimensional
grid called the map. Every node of the map is connected to a
common set of input. Any activity pattern on the produces a
variable pattern of activity in the nodes, and the map is
updated so that neighboring cells on the map are made more
similar to the most active node (best fit to data item), which
in turn leads to topological clustering on the map, reflecting
recurring similarities among data items. The result is a map
where similar inputs are placed close to each other, and
dissimilar items further apart.
Since there is no need for a priori classifications of the
input (the learning algorithm of the map is unsupervised) the
map is a useful tool in exploratory data analysis and
visualization. The map itself, however, provides to
conceptual interpretation of the clusters nor a discrete
partitioning of the inputs into classes of similar items, which
need to be provided by the researcher interpreting the map.
Results & Discussion
Ontological Commitments
The first set of questions probed the basic ontological
commitments of the students. Overall, it seemed that about
70% of our Finnish upper secondary school students were
1676
dualists of some kin. For instance, when they were assessed
the claim “The mind is material”, 73% answered “no”, and
when they were asked if the mind is immaterial” 69%
answered “yes”. However, this does not imply that 70% of
the students would have answered the question on the basis
of a coherent substance dualism. When they were asked if
the human mind is an immaterial entity, only 50% of
students found it to be so. How can these percentages be
reconciled? It may be that either students have difficulties to
understand the notion “entity” in a mentalistic context
(fragmentation), or they may alternatively think that the
mind is a sort of event or a process (ontological theory)2.
A hierarchical cluster analysis using Ward’s method on
the first set of questions suggested the students might be
divided into four categories. When the pattern of responses
to these questions was analyzed qualitatively, a four
category typology of responses types could be constructed,
based on three diagnostic subsets of the questions (cf. Table
1). Subset 1 (questions 2, 9, 10 & 18) asked if the mind is
immaterial rather than material, Subset 2 (14 & 37) asked if
the mind has a material basis, and Subset 3 (15, 20, 22 &
35) asked if the mind could exist without a material basis.
One group (A) mostly denies the immaterial and
independent nature of the mind. These students are
apparently some sort of materialists. Another group (D) has
a diametrically opposing view: they consider the mind to be
immaterial and, although many say the mind has a material
basis, to be at least in principle independent of it. They
therefore seem to be traditional dualists. A third group (C)
considers the mind to be immaterial, but nevertheless
dependent on a material basis for its existence. These
students may be some kinds of dualists, or “emergentist”
materialists. Finally, there is a fourth group (B), for whom
we struggled to find any straightforward interpretation. It
may be either that these students lack a consistent outlook
on the ontology of the mind, or else the items on the
questionnaire fail to probe it well enough. These alternatives
are left as open questions for future research.
When the results of this set of questions were fed into a
self-organizing map we could identify these types with
clusters of students with similar overall response patterns
(fig. 1). Looking at the map, we see that of the four groups
of students group D (“dualists”) form a well-defined cluster
in the lower right corner of the map. Group A
(“materialists”) are located at the opposite (upper left)
corner, and are also fairly nicely clustered. Group C
(“emergentists”) is located in the center, somewhat nearer
the materialists than the dualists, and is more diffuse.
Finally, members of group B are much more scattered
throughout the map. This is the group for which we failed to
find any coherent content-description. (It may be that with
more subjects/items this group would divide into several).
2 Students also seemed quite sensitive to the exact wording of
the questions. For example, most students answered that a mind
has a material basis, viz. the brain (91%), but when they were
asked if the mind has a material basis but the brain was not
mentioned, only 75% answered “yes”.
Table 1:
Classification of the students according to ontological
commitments and the groups’ average response percentages
to questions probing ontological commitments (see text).
Group Subset 1 Subset 2 Subset 3
A 17,3 % 100% 3,9% n=13
B 64% 75% 39% n=16
C 79,6% 91% 6,8% n=22
D 91,2% 62% 95,6% n =17
All 66,9% 81,65% 36% n=68
Hypothetical reasoning
The second and third sets of questions probed intuitions on
the properties and competences of an immaterial mind.
Since the aim of the questions was to reveal the structure of
conceptual organization of students’ beliefs about material
and non-material mental attributions, the structure of these
questions was hypothetical: the students were asked to
assume that if there existed an immaterial mind,
independent of physical body, could it have such and such
spatial or temporal attributes or such and such perceptual or
cognitive abilities.
Looking at the responses in light of the classification
discussed above, we could not establish any clear pattern
between the four types of ontological commitments and the
responses to hypothetical questions. Forcing the students to
reason hypothetically therefore seems to tap into different
similarities and differences than those that manifest in
explicit ontological claims.
For example, each of the four groups contained subjects
who associated spatial attributes with an immaterial mind
(about 60% of the students, overall) and subjects who did
not. For example, when asked if a an immaterial mind
separate from the body nevertheless is located somewhere,
62% of the students answer “yes”, while 56% thinks that if
reincarnation of the soul were possible the soul would be
located somewhere “in between the bodies”, as it were.
When asked more closely about this transition, 46-59% -
depending on the question - of the students said that the
mind would occupy intermediate locations between the
locations of the old and the new body (questions 53 & 59).
This suggests that the students do indeed consider the
location of a reincarnating soul as concrete physical or
geographical location.
When the responses to hypothetical questions were fed
into a SOM, thematically some thematically related items
were found to form clusters on the map, others not.
Questions relating to the spatial motion of immaterial mind
(questions 57, 50, 60, 13, 26, 7, 28, 53, 59, 57, 55, 61, 16,
56, 63) are all located around the middle section of the map,
but do not form a neat cluster (their distances from each
other are large, as indicated by shading).
1677
65F_A
31F_A
15F_A
17M_A 52F_A
24M_C
21F_B
12M_B 13F_B 61M_B
45M_B
32F_B
28M_B
11M_ A 38 M_ A 63M_B 67F_B 57M_B
48M_A
34M_A
19F_A
8F_A
16M_C 40M_C 62F_C 68M_D 35F_D
6M_D
56F_C 59M_C
18F_C 9M_B 42M_D
43F_C
33M_A
49F_C
20F_C
39M_C
30M_C
23M_C
7M_C
2M_C
53M_C 55F_B 51F_D
3M_D
44F_D
41M_D
37F_D
36F_A 5M_C 64F_C 60M_D 1F_D
22F_B
4M_B 27M_C 47M_C
26F_C
29M_B
10M_C 25M_C 54F_D 14M_D
66F_D
58F_D
50F_D
46M_D
Figure 1: Self-organizing map of the students.
The map is based on the proximity of response patterns on questions probing ontological commitments. Darker square
indicates longer distance in the input vector space. Students numbered 1-68 in no particular order. M = male, F = female.
Labels A-D indicate membership in categories classified by a cluster analysis on the same data.
On the other hand, responses to the capacities an
immaterial mind cluster much more tightly – but not into
a single cluster.
Approximately 40% of students seemed to think that
immaterial Cartesian minds, if such existed, could have
abilities that are dependent on the sensory or perceptual
processes (Table 2). For instance, 44% assumed a mind
without a body could see colors (question 23), 40%
thought it could see geometrical shapes (36), and
according to 40% it could it hear sounds. Yet only 19%
said the mind could feel heat (58), and only 18% said it
could experience tastes. Looking at the distribution of the
items related to perceptual and cognitive abilities of a
hypothetical immaterial mind, we find the questions
related to visual and auditory perception (39, 36, 23) in
one tight cluster, questions related to feeling warmth and
tastes (58 & 62) a little further, separated by questions
asking whether alcohol intoxicates the body, the mind, or
both (33, 30, 32) – with questions related to higher
cognitive function (decision making 44, 51, thought and
mental calculation, 41, 21). Apparently sensation – even
in immaterial minds - is conceived differently from
cognition. (More “bodily”, perhaps).
34 57 60
50
26
13
38
52
48 53 742
41 59 28
31
36
39
47
51
46
61
55 40 23
16 33 30
45
44
43
29
21
63
56
62
58 32
Figure 2.
SOM map of the hypothetical questions. Label refers to
question item on questionnaire. Nearby items have similar
response patterns across the sample.
Surprisingly, when higher cognitive (non-perceptual)
abilities of an immaterial mind, such as an ability to make
inferences, is considered, only about 60% of the respondents
are willing to attribute such properties (see Table 2).
According to some 40% of the respondents, Cartesian minds
do not have characteristics Descartes considered essential to
1678
the soul, a difference of opinion on one of the main
premises of a key argument in any introductory philosophy
course.
The overall structure of the map suggests that the students
do not share a clear and coherent set of beliefs on the
spatiotemporal attributes of an immaterial mind, whereas in
the case of sensory and cognitive capacities they are quite
consistent. The question of internal coherence of this
recurring set of belief remains an open question, however.
The SOM map cannot address this question directly –
however, it does show that if the students are incoherent,
they are consistently incoherent in the same way.
Table 2: The abilities of an immaterial mind
Immaterial mind is able to % of students
36. See geometrical shapes 40
23. See colors 44
39.Hear sounds 40
21.Make inferences & perform
calculations
41.Count the number of ”objects of
thought”
44.Make decisions
58. Feel heat
62. Taste tastes
59
62
65
19
18
Conclusion
We have presented a framework and an application for
discipline-based research on philosophy learning. In our
empirical design we focused on the initial state of the
learning process – the novices’ conceptions about
philosophical topics students hold when they are first
exposed to academic philosophy and the tradition of western
philosophy. The aim of the study was not only to capture
some features of philosophy novices’ conceptual
organization but also to give us a hint of the degree of their
belief system’s consistency coherence – whether students
are all alike in their beliefs, and whether beliefs that should
go together do go together. We believe this might offer a
fertile and fruitful framework for investigating the transition
process from the novice to the expert, as establishing
conceptual coherence is precisely what learning philosophy
is all about.
Acknowledgements
Eero Salmenkivi and especially Gualtiero Piccinini for
commenting an earlier draft on this paper.
References
Bechtel, W. & Richardson, R. (1993). Discovering
Complexity, Decomposition and Localization as
Strategies in Scientific Research. New Jersey: Princeton
University Press.
Carey, S. (1985). Conceptual Change in Childhood.
Cambridge, MA: MIT Press.
Chi, M., T.H. (1992). Conceptual Change within and across
Ontological Categories: Examples from Learning an
Discovery in Science. In R. Giere (Ed.), Cognitive Models
of Science. Minneapolis, MN: University of Minnesota
Press.
Chi, M. & Slotta, J. (1993). The Ontological Coherence of
Intuitive Physics. Cognition & Instruction. 10 (2 & 3),
249-260.
Chi, M. T H., Slotta, J. D., & de Leeuw, N. (1994). From
Things to Processes: A theory of Conceptual Change for
Learning Science Concepts. Learning and Instruction, 4,
27-43.
Craver, C. (2006). When Mechanistic Models Explain.
Synthese, 153, 355–376.
Cummins, R. (1983). The nature of psychological
explanation. Cambridge, MA: Bradford/MIT Press.
DiSessa, A. A. (1988). Knowledge In Pieces. In G. Forman
& P. Pufall (Eds.), Constructivism in the Computer Age
(pp. 49-70). Hillsdale, NJ: Erlbaum.
DiSessa, A. A. (1993). Toward an Epistemology of Physics.
Cognition and Instruction, 10, (2 & 3), 105-225.
Limón, M., & Carretero, M. (1999). Conflicting Data and
Conceptual Change in History Experts. In W. Schnotz, S.
Vosniadou & M. Carretero (Eds.), New Perspectives on
Conceptual Change (pp. 137-160). Amsterdam:
Pergamon.
Murphy, G.L. & Medin, D.L. (1985). “The Role of Theories
in Conceptual Coherence“, Psychological Review, 92 (3),
289-316, 1985.
Kohonen,T. (2001): Self-Organizing Maps. 3rd edition,
Springer, 2001.
McCloskey, M. (1983). Naive Theories of Motion. In D.
Gentner & A. N. Stevens (Eds.), Mental Models (pp. 299-
324). Hillsdale, NJ: Erlbaum.
Samarapungavan, A. & Wiers, R. W. (1997). Children´s
Thoughts on the Origin of Species: A study of
Explanatory Coherence. Cognitive Science, 21 (2), 147-
177.
Slotta, J. D., Chi, M., T.H., & Joram, E. (1995). Assessing
Students Misclassifications of Physics Concepts: An
Ontological Basis for Conceptual Change. Cognition and
Instruction, 13 (3), 373-400.
Strike, K.A. & Posner, G.J. (1982). ”Conceptual Change
and Science Teaching”. European Journal of Science
Education, 4 (3), pp. 231-240.
Thagard, P., & Verbeurgt, K. (1998). Coherence as
Constraint Satisfaction. Cognitive Science, 22 (1), 1-24.
Thagard, P., Eliasmith, C., Rusnock, P., & Shelley, C.
(2002). Knowledge and Coherence. In R. Elio (Ed.),
Common Sense, Reasoning and Rationality (pp. 104-131).
New York: Oxford University Press.
Vosniadou, S., & Brewer, W. F. (1992). Mental Models of
the Earth: A Study of Conceptual Change in Childhood.
Cognitive Psychology, 24, 535-585.Woodward, J. (2003):
Making Things Happen: a Theory of Causal Explanation,
Oxford: Oxford University Press.
1679
