Abstracts for ConChaMo 3

A conceptual and referential clarification of the category of physical quantities in the perspective of science education research

Strömdahl, Helge
Linköping University, Sweden

A lot of research in science education, during decades, has been devoted to the learning and teaching problems of concepts like force, energy, heat, electric current. What is less noticed is that these concepts are belonging to the scientific category of physical quantities. Moreover the nature of these concepts, e.g. their reference is seldom considered. Chi, et al. (e.g.1992;1994;1995) has touched upon this issue by viewing conceptual change as a question of change of ontological categories. Especially the change between the material category to the category of phenomena is essential in their approach. For instance a lay concept of electric current as some material fluid should be correctly recognized in science as a constraint-based phenomena, indicating a need of conceptual change in that direction. However electric current is a physical quantity, belonging to the category of physical quantities, and as a physical quantity by definition is a measurable property of a phenomenon or object not the phenomenon or object itself. The category of physical quantities is not accounted for by e.g. Chi, et al. in their ontological category approach to conceptual change. I will, pro primo argue for the recognition of the category of physical quantities and pro secondo from a semantic/semiotic analysis approach argue that the identification of the referent of a physical quantity is a factor to be recognized in conceptual learning and conceptual change. However, it is a challenge to discern the referents of physical quantities, an issue that will be explored and discussed in my presentation.

Language, Metaphor and Ontology in Conceptual Change

Amin, Tamer
American University of Beirut

Use of verbal predicates has been seen by conceptual change researchers as an important indicator of the ontological classification of concepts by experts and novices. Meanwhile, research on conceptual metaphor in cognitive linguistics has revealed extensive patterns of metaphorical language use reflecting systematic mappings between conceptual domains. Such mappings have been shown to be pervasive in both everyday and technical (scientific and mathematical) language use. In this presentation, I summarize findings from a number of studies examining metaphorical language use in scientific texts as well as expert and novice reasoning and problem-­‐solving. These studies used analyses of patterns in language use to infer the use of experience-­‐based knowledge gestalts (e.g. containment, object possession and movement, force dynamics) to construe and reason about abstract concepts in thermodynamics (e.g. energy and entropy). While science concept learning does seem to involve some dramatic conceptual changes, the continued use of experience-­‐based knowledge gestalts in advanced scientific conceptualization and reasoning is not readily assimilated to a strict ontological reclassification view of conceptual change. The presentation ends by discussing the implications of the findings presented for how we might theorize relationships between language, metaphor and ontology in conceptual change.

Considering Students Conceptions as Complex Dynamic Systems

Brown, David E.
University of Illinois at Urbana-Champaign

With a student's conception considered as a complex dynamic system or CDS (i.e., as like a dynamic ecosystem rather than as like a* regular thing*, such as a rock), aspects of students' conceptions and conceptual change, which are surprising if these conceptions are considered as regular entities, become expected if students' conceptions are considered as CDS's. First, with CDS's, at times strong influences can lead to little change (strong stabilities or *attractors* develop that are affected little by external influences), which predicts the kind of robustness often seen with students' conceptions. Second, CDS's are emergent and evolving rather than static. Any identifiable systematicities in student thought would then be the result of dynamic emergence from the complex dynamic conceptual system. Such emergent structures can be fleeting or highly stable, accounting for both strong coherence and significant contextuality in students' conceptions, both of which have been seen in numerous studies. Finally, CDS's are embedded in and embed other CDS's. We would therefore expect interactions among various levels of complexity that impinge on students' conceptions: subconceptual, conceptual, metaconceptual, discursive, sociocultural, etc. A CDS view encourages consideration of such embeddedness, without a reification of one level to the exclusion of others. Such a multidimensional perspective, a natural outgrowth of a CDS view, is increasingly seen as important in considerations of student conceptual change.

Concepts, Ontological v Categories and the *No-Overlap* principle.

Tobin, Emma
University College London

Kuhn´s no-overlap principle (Kuhn 2000) precludes the cross-classification of objects into different natural kinds within a single taxonomy. Cases of conceptual change break the no-overlap principle (Kuhn 2000a: 92­6) because theories separated by a revolution cross-classify the same things into mutually exclusive natural kind categories. In this paper, I will discuss some of the major ontological theories of natural kinds and relate them to cases of conceptual change. I will claim that the Kuhnian no-overlap principle places too strong a constraint on conceptual change and does not cohere with some obvious examples from scientific practice. I will look at some case studies to examine (a) how natural kind categories change/evolve within a single taxonomy and (b) how natural kind categories change across a scientific revolution. Finally, I will examine the role of concepts in both.

Conceptual change and ontological issues - heat, entropy and microstates

Jesper Haglund
Linköping University, Sweden

We present findings from four empirical studies of students in different ages interacting with thermal phenomena, and reflect on how ontological issues are involved in understanding of the concepts of 'heat', 'entropy' and 'microstates'.

7- to 8-year-olds were found to identify heat with warm matter, a substance view, in heat convection, but to conceptualise heat conduction between solids more in terms of 'contagious action'. 12- to 13-year-olds were found to be reluctant to talk of 'heat' as a noun involved in heat conduction, and had therefore difficulties seeing heat flowing in a metal knife when it was warmed by their thumbs, even when aided by a heat camera. In all, a process view of heat conduction typically seems to predate a substance view.

The issue of the ontological classification of entropy and the proposed idea to introduce entropy as 'the everyday conception of heat' are briefly brought up. Empirically, preservice physics teachers were wrongly found to believe that the entropy should increase in reversible, adiabatic expansion of an ideal gas. This may be explained in terms of an ontological commitment to microstates as locations in physical space, rather than as metaphorical locations in phase space, taking into account the energy distribution. In contrast, physical chemistry PhD students productively conceptualised microstates, on the one hand, as objects to be placed in a box and, on the other, as locations in which particles could be placed, depending on the quantitative modelling involved in problem-solving.

In conclusion, we embrace the idea of conceptual change as striving to “extend our repertoire of ideas” (Caravita & Halldén, 1993), rather than replacement or abandonment of prior conceptualisations.

Conceptual knowledge and its change seen as a complex system

Ismo T. Koponen

The notion of complexity has been applied in several ways to describe the problems of learning, instruction and conceptual change. The viewpoint of complex systems fit quite naturally on such field of problems, where diverse interaction of different “agents” takes place to produce learning. In one end of the spectrum the interactions are individual relations, learners interact with other learners, instructors, and different material and artifacts that belong to the learning environment. In the other end of the spectrum, the interactions are between elements and structures of individual’s cognitive system, e.g. concepts and conceptual structures. In very broad and general terms the science of complexity has provided a suggestive terminology to speak about these types of relations, where “agents” interact and as its consequence, collective or emergent new “states” are appearing. In level of socio-dynamic behavior such states can be stable groups and information flows between the groups, in level of cognition, they can be new conceptual structures. However, in order to take a step further and to progress beyond the stage of renaming the old and known phenomena by using the vocabulary of complexity, we must try to identify the basic agents and their basic relationships and then to build idealized models to describe the dynamics of such systems. This kind of idealization and simplification is actually in the core of science of complex system. In a sense, it is the simplex approach on complexity.

In case of conceptual change one promising line of approach might be identifying first some elemental conceptual schemes, and the trying to build a model system which operates on basis of mutual relationships between these elemental schemes. What makes such a line of attack a promising one, is that there is already existing background to identify such elemental schemes, possibly in form of p-prims or simple framework theories. Also, the idea of stabilizing interaction patterns between conceptual schemes has been around quite a long time, in rather well advanced idea of coordination class. This kind of approach, on the other hand, yields to description in a form of a network, where different elements represent the nodes and the connections the relationships. Of course, the network is only an idealized representation, a kind of map cartographic the different possibilities of interaction patterns. Its main advantage is not actually getting in the picture as much as can be achieved - an attempt easily resulting to messiness instead of complexity – but rather to leave out as much as possible and retain only the essential connection and elements – like in a subway map of a large city or in an organization scheme of large transport system.

In this presentation I suggest one possibility to build a model of conceptual change from the vantage point of complex systems, and to represent such a system as a dynamic network. The example is about the differentiation of two closely related physics concepts, electric current and voltage. The modeling of the process of conceptual change is conceptualized so that it is grounded on empirical evidence, gained through traditional research methodology based on interviews, but the way to look the is strongly guided by theoretical conceptions of concepts as structures, where concepts are connected on the other hand to characterizing (but not defining) attributes, and on the other hand to theoretical-type constraining models and explanatory models, which act as a rules how concepts can be related and projected on real phenomena. These elements form a system, which changes dynamically when context changes, but which can stabilize on certain states, identified as stable learning outcome. The model is still preliminary, but it already displays what it might in practice and in level of modeling if the idea of conceptual knowledge and it changes is seen from a viewpoint of complex system.

The Challenges of Understanding Emergent Phenomena: Implications for Science Education

Jim Slotta

As a PhD student in 1990-95, I worked with Professor Michelene Chi at the University of Pittsburgh, where my thesis study provided the first empirical support for her (then) new theoretical model of conceptual change. In short: people have trouble understanding certain kinds of concepts because (a) they have pre-existing conceptualizations which are ontologically compatible with the scientifically normative view, and (b) they do not have a well established "target" ontology into which they could map new instances. My thesis work developed an instructional approach that sought to address both of those problems, and Dr. Chi has subsequently performed a range of studies that explored such learning designs. Since 2005, I have been developing a new model of collective inquiry (i.e., where the entire classroom is considered as the "unit" of analysis), that allows for complex, sustained interactions over many weeks of curriculum. I have recently applied this model, called Knowledge Community and Inquiry (KCI) in a series of design-oriented studies that investigate student learning of emergent phenomena such as food web relationships and biological evolution. My presentation will connect the psychological foundations of Dr. Chi's theory with my own model and instructional designs. In particular, I will focus on an epistemic shift toward collective inquiry, and a possible role of embodied cognition.

What computational analyses reveal about concepts and conceptual change

Bruce Sherin

In recent research, I have been using computational methods to analyze transcripts of student interviews. The computational methods I draw upon are various "bag of words" techniques from computational linguistics. These techniques merely "count words; they pay little attention to the order in which words are spoken. Although these computational methods use very simple techniques and limited data, they nonetheless are able to replicate many features of detailed qualitative analysis by my research team.

For this talk, I will discuss my work with a corpus of interview data in which students were asked to explain the Earth's seasons. The input to the computational analyses are transcripts of these interviews that include only the words spoken by the participants. The output has two parts: (1) looking across the entirety of the corpus, the computational analyses produce a set of concept codes; (2) these concept codes are then used to produce analyses of individual interviews which show how the different concepts appear, and in what configurations over the course of the interview.

The success of these methods poses important questions for our theories of conceptual change, as well as for the relationship of those theories to our data and analytic methods. How is it possible that simple methods --- methods that have no access to student gestures and drawing, and that even ignore word order --- can nonetheless replicate careful qualitative work by human analysts?