ZDM Mathematics Education (2008) 40:189199 DOI 10.

1007/s11858-008-0083-2

ORIGINAL ARTICLE

On the inuence of theory on research in mathematics education: the case of teaching and learning limits of functions
Christer Bergsten

Accepted: 1 March 2008 / Published online: 18 March 2008 FIZ Karlsruhe 2008

Abstract After an introduction on approaches, research frameworks and theories in mathematics education research, three didactical research studies on limits of functions with different research frameworks are analysed and compared with respect to their theoretical perspectives. It is shown how a chosen research framework denes the world in which the research lives, pointing to the difcult but necessary task to compare research results within a common eld of study but conducted within different frameworks.

1 Introduction It is generally acknowledged that results from didactical research, as any other research on human behaviour in social settings, depends heavily on the underlying basic assumptions, general approach, and theories and methods used. For example, Artigue, Batanero and Kent (2007) state that favouring certain approaches shapes the problematics and methodology of the research, and through these the kind of results that one can assess and the way they will be expressed (p. 1030). One may thus ask, for a particular study, what factors inuence the choice of a specic research framework, and what consequences this choice entails. Such considerations seem crucial when evaluating not only the validity of the research design but also the scope and consequences of the specic research results, for
C. Bergsten (&) Department of Mathematics, pings Universitet, Linko ping, Sweden 58183 Linko e-mail: chber@mai.liu.se

example to be able to judge its relevance for different educational contexts. They may also be useful considering the diversity of theoretical tools and frameworks in present day mathematics education. According to Silver and Herbst (2007, p. 43), the increase of this diversity during the last decades has been caused by the turn away from quantitative toward qualitative research methods, as well as new methods for experimental research. Other arguments given to account for this diversity are the diverging, epistemological perspectives about what constitutes mathematical knowledge and the strong inuence of cultural, social, and political forces on mathematics education (Sriraman and English 2005, p. 452). Indeed, it was concluded in Sierpinska and Lerman (1996) that for mathematics educators, epistemologies of mathematics and assumptions about the epistemic subject cannot be divorced (p. 868), and they also noted at that time that There is much debate within the international community of mathematics educators about theoretical approaches and their underlying epistemological issues (p. 867). One way to shed light on how frameworks, theoretical tools and methods shape the design and product of didactical research is to compare different kinds of studies on the same or similar educational phenomena with regard to these dimensions. Such comparisons may also have other more pragmatic objectives besides the academic quest for understanding the roles of different components in a research process, such as facilitating an integration or a network of the outcomes of the studies to contribute to a progression of our didactical knowledge of these phenomena. Taking isolated research results at face value, without relating them to the conditions and constraints of the research processes behind them, provides no criteria or bases for relating them to other seemingly contradictory or similar results, and is therefore of limited value in the

123

190

C. Bergsten

construction of a coherent knowledge base in an educational science. After a short introduction on research frameworks and the concept and role of theory in mathematics education, this paper then goes into more detail looking at didactical research on a specic mathematical notion, limits of functions, often referred to as difcult for students to learn or understand (Mamona-Downs 2001). I will give a short overview of some approaches and perspectives used in educational research on limits, and then compare more closely three specic studies, representing different research frameworks, with a focus on their theoretical underpinnings and claims. In doing this, I will consider the following question: How does a theoretical basis adopted for a study inuence the nature of the purpose, questions, methods, evidence, conclusions, and implications of the study? A short remark on terminology is needed here. In the next section, I will elaborate on the terms approaches to research and research frameworks. However, in the formulation of the question above, the term theoretical basis is used to designate the overall purposive positioning of the research, including its approach and research framework (as described in the next section). The comparison between the three studies also serves as a case study and a method to investigate the question above, using the analytical tools presented in the next section. The aim of this enterprise is twofoldto better understand the consequences of the adoption of a specic theoretical basis for a research study, and to provide a basis for discussing how such a comparison can facilitate an integration or a network of the outcomes of the studies, in order to contribute to a progression of our didactical knowledge of these phenomena.

2 Research frameworks and theories In Lester (2005) reasons are given for why educational research needs to be pursued within a scaffolding framework. A framework is here seen as a basic structure of the ideas (i.e. abstractions and relationships) that serve as the basis for a phenomenon that is to be investigated (p. 458), representing its relevant features as determined by the adopted research perspective, and serving as a viewpoint to conceptualise and guide the research. A research framework thus provides a structure for conceptualising and designing research studies, including the nature of research questions and concepts used and how to make sense of data, allowing to transcend common sense (p. 458). According to Eisenhart (1991) three kinds of research frameworks can be identied, that is a theoretical, a practical, and a conceptual framework. In relation to

theoretical frameworks, Lester (2005) argues that although making the choice of conforming to a particular theory has the advantages of facilitating communication, encouraging systematic research programs, and demonstrating progress (p. 459), it also has serious shortcomings, such as prompting explanations by decree rather than evidence, making data travel to serve the theory, offering weak links to everyday practice, and limiting validation by triangulation. Practical frameworks, based on accumulated experiences and what works, may suffer from limitations caused by norms and narrow insider perspectives. The focus of a conceptual framework is more on justication than on explanation but still based on previous research. Instead of relying on one particular overarching theory as in the case of a theoretical framework, it is built from an array of current and possibly far-ranging sources, and can be based on different theories and various aspects of practitioner knowledge, depending on what the researcher can argue will be relevant and important to address about a research problem (Lester 2005, p. 460). The validity for the chosen framework is context dependent, which is its strength considering the practical implications of the research. Lester thus pragmatically argues with an emphasis on justication, the purpose of research to answer the why questions, that we should focus our efforts on using smaller, more focused theories and models of teaching, learning and development (p. 460). The notion of a conceptual research framework relates to the idea of a networking strategy for dealing with the increased diversity of theories within mathematics education (Bikner-Ahsbahs and Prediger 2006). In a similar vein, when discussing four broad theoretical perspectives, Cobb (2007) uses a bricolage metaphor: rather than adhering to one theoretical perspective, we act as bricoleurs by adopting ideas from a range of theoretical resources (p. 29; cf. Gravemeijer 1994, p. 447). Since the development of such perspectives has been initiated by other purposes than those driving mathematics education research, they may serve as sources of ideas that we can modify for our purposes as mathematics educators (Cobb 2007, p. 29). In line with this view, the researcher should not only make explicit the choice of theoretical frames used but also the justication for this choice. Meanings and different uses of the word theory have been discussed for example in Mason and Waywood (1996) and Silver and Herbst (2007). Niss (2007a, b) notes that though the notion of theory is essential for mathematics education research, and often used, a denition of the key term theory is seldom or never explicitly given. He goes on to offer such a description of this notion, stating that a theory is an organised network of concepts and claims about a domain, where the concepts are linked in a connected hierarchy and claims are either basic hypotheses

123

On the inuence of theory on research in mathematics education

191

taken as fundamental, or obtained from these by means of formal or material derivation. To be a theory this network is also required to be stable, coherent, and consistent. Niss (2007a, b) separates the purpose of using theory and its role in research. In the former category he lists explanation, prediction, guidance for action or behaviour, a structured set of lenses to study phenomena, a safeguard against unscientic approaches, and protection against attacks from sceptics in other disciplines. Concerning the role of theory he mentions providing an overarching framework, organising observations/interpretations of related phenomena into a coherent whole, terminology, and research methodology. He also adds that the inclusion of theory in general is needed for publication. Similar points are raised in Silver and Herbst (2007), who view the different roles of theory as mediators of connections between the vertices of what they call the scholarship triangle (see Fig. 1). Within each of these connections, Silver and Herbst describe several different roles that theory could play. For example, for the link between research and problems, one of the roles that theory serves is as the means to transform a commonsensical problem into a researchable problem (p. 50). Mathematics education is characterised by its double nature (Niss 1999), with both a descriptive purpose, aimed at increased understanding of the phenomena studied, and a normative purpose, aimed at developing instructional design. In discussing the role of theory in research, the dynamic model presented in Lester (2005) takes this double nature into account (see Fig. 2). The primary outcome of research may be to increase understanding of a specic phenomenon or to improve practice, a goal pursued along different possible pathways of pure, basic, applied, or developmental research. Such outcome then can serve as starting point for a new research process. One note to make is that it is the dynamics of this cycle that may establish connections between understandings and products.

Improved understanding

Improved products

Pure, basic research

Useinspired, basic research

Pure, applied research & development

Existing understanding

Existing products

Fig. 2 A dynamic model of educational research (Lester 2005, p. 465)

Problems

Research

Practices

Fig. 1 The scholarship triangle (Silver and Herbst 2007, p. 44)

A research question in education often has its origin in a commonsensical problem, that is a general reection expressed in everyday language that, for example, some mathematical concept is difcult to understand for some group of students, or that they often make a certain kind of mistake on specic types of problems (cf. Prediger in this issue). The problem identied may then be approached from different angles or perspectives in order to describe it for research purposes and formulate a research question. The decision on what approach to apply rests not only on issues such as the character of the problem and the empirical context, or the aim of the research in terms of for example generality and practical applicability of results. Critical are also the backgrounds, interests, and theoretical orientations of the researchers, including their explicit or implicit assumptions on mathematical knowledge and learning in educational settings. One may identify at least three different general approaches used in research on mathematics education: an epistemological, a cognitive, and a social approach. In an epistemological approach, focus is on the structure and use of mathematical knowledge and its diffusion in educational institutions. Within the cognitive approach the research interest is focused on the mental structures and thinking processes involved in learning, understanding and doing mathematics, including meta-cognitive dimensions. Taking the social and situational context of the learner into account, for example a classroom perspective, or involving more broad social and cultural factors on mathematics education, a social approach is used. While acknowledging the fact that, for example, a study with an epistemological approach needs to consider social and cultural dimensions of knowledge, or that an epistemological analysis of the object of learning may be used within a cognitive approach, this distinction is made here to identify the main approach or focus/interest of the study. The approach chosen largely determines in which domains descriptions

123

192

C. Bergsten

and explanations of observed educational phenomena are sought. What is considered a problem within one approach may be viewed as a symptom of another kind of problem within another approach. The kind of theoretical basis to frame the study is also suggested but not implied by the approach taken. Sierpinska and Lerman (1996), while discussing epistemologies in mathematics education, use the terms view and approach about the main such epistemologies they identify: constructivism, socio-cultural views, interactionist views, an anthropological approach, and approaches based on epistemologies of meaning. Recently, in the same vein, Artigue et al. (2007) discuss cognitive, epistemological, and more global approaches used in research on post-secondary level mathematics education. While using similar categories, the notion of approach presented above is to be understood as a more informal inclination by the researcher to interpret an observed commonsensical problem, be it more or less biased in the researchers main theoretical orientation or interests. When coping with the diversity of theoretical perspectives used in mathematics education, Cobb (2007) suggests to focus on types of questions asked, the nature of the phenomena under study, and the forms of knowledge produced rather than trying to establish direct links between theory and instructional practice. He compares four broad theoretical perspectives used in mathematics education research, experimental psychology, cognitive psychology, sociocultural theory, and distributed cognition. As a basis for contrasting these perspectives Cobb sets up two criteria: how each perspective characterizes individuals and the potential of each to contribute to our understanding of learning processes and the means of supporting their realization (p. 28). He stresses how the choice of these criteria is relative to a view of mathematics education as a design science and how their use involves interpretation and judgement (p. 28). The cognitive and social approaches discussed above are related to at least the last three of these four perspectives, making the outcome of Cobbs contrasting comparisons, expressed in terms of characterization of the individual, usefulness, and limitations (see p. 28, Table 1.1), useful also for the comparisons of individual studies based on these approaches. However, epistemologically oriented theoretical perspectives, such as the theory of didactical situations or the anthropological theory of didactics (see Bosch and Gascon 2006), are not discussed by Cobb. In the following, I will be using the terms research framework, theoretical framework, and conceptual framework in line with Eisenhart (1991) and Lester (2005), and the term approach to refer to one of the categories epistemological, cognitive, and social, as in the discussion above.

3 Research on the mathematical notion of limit Overviews of research on limits are found in Cornu (1991) and in Harel and Trgalova (1996). Cognitive approaches have dominated this research, identifying the critical roles played by conceptions of innity, quantication, visualization, concept images, the dialectic between processes and objects, and between intuition and formalism, conceptual metaphors and image schemata, and students beliefs about mathematics and their role as learners. Epistemological approaches have discussed historicalphilosophical aspects of the mathematical ideas involved in the limit concept (e.g. Burn 2005), epistemological obstacles (e.g. Cornu 1991), or contrasted mathematical and didactical organi sations of knowledge observed in classrooms (e.g. Barbe et al. 2005). Some examples given below show how different combinations of main purposes, approaches, and research frameworks are used in recent research. Juter (2006) applies a cognitive approach, using a conceptual framework with a focus on concept images and the three worlds of Tall (2004) to investigate Swedish university students understanding of limits. Her study conrms the image of limits as a problematic area, but that students often tend to overestimate their own abilities as compared to their achievements. While the aim of Juters study was improved understanding, in terms of Fig. 2, the goal for Przenioslo (2005) was improved products for introducing the limit concept. Based on a similar approach as Juter, using a conceptual framework, she outlines an instructional design based on a didactical tool to enable students to develop conceptions that are closer and closer to the meaning of the concept of limit of a sequence (p. 90). By taking the characteristics of the object of knowledge for granted, this study is not classied as adopting an epistemological approach. Mamona-Downs (2001) also aims at developing a teaching/learning practice by making tacit intuitive views visible and conscious, applying a cognitive approach paired with an epistemological analysis of the pre-given mathematical denition of the concept of limit. Bergsten (2006) applies an epistemological approach and a theoretical framework to analyse university students work on limit tasks with a focus on understanding the role of algebra in their mathematical reasoning. In the next section, three studies are described in more detail in order to discuss the consequences of using particular approaches and research frameworks. To reect that cognitive approaches dominate the research on the specic content area of limits of functions, two such studies were chosen but with different kinds of research frameworks, that is a theoretical and a conceptual framework. To contrast approaches, a study with an epistemological approach was also included in the analysis, using a theoretical framework. There are thus two studies with the

123

On the inuence of theory on research in mathematics education

193

same approach but different kinds of research frameworks, while two differ in approaches but are both conducted within a theoretical framework. After presenting my interpretation of these studies, a comparison between them is made structured by the research question stated in the Sect. 1, and using the analytical tools discussed in the previous section. 3.1 Cognitive schemas An example of a cognitive approach is found in Cottrill et al. (1996), where the theoretical framework used is explicitly stated in the paper as the APOS theory, based on Piagets constructivism. The focus is on students understanding of the limit concept, and after acknowledging student difculties to understand this concept, the stated purpose is to apply our theoretical perspective, our own mathematical knowledge, and our analyses of observations of students studying limits to develop a description, or genetic decomposition, of how the limit concept can be learned (p. 167). This tool is based on the APOS theory, in particular how it treats the reconciliation of the dichotomy between dynamic or process conceptions of limits and static or formal conceptions (pp. 167168). The perspective is based on the following statement about mathematical knowledge (p. 171): Mathematical knowledge is an individuals tendency to respond, in a social context, to a perceived problem situation by constructing, re-constructing, and organising, in her or his mind, mathematical processes and objects with which to deal with the situation. The acronym APOS designates the sequence actionprocess-object-schema and by this focus the theory considers how learners, when reecting on actions on (mathematical) objects, come into control of the processes involved in the transformation of objects. By internalization these processes can be encapsulated into new objects. This triad establishes a mental schema, which may be used as a basis for a new action sequence. The APOS theory thus depicts a mechanism that may be envisioned as a spiralling of action, process, and object within expanding schemas (p. 173). The adopted theoretical basis for this study is mirrored in the terminology used, such as the frequent terms construct and schema, as in the coordinated process schema is difcult in itself and not every student can construct it immediately (p. 174). The conclusion is an instructional design focusing on getting students to make specic mental constructions (p. 169) of importance for understanding the limit concept. The research method is a cyclic process, where a genetic decomposition is developed by an epistemological analysis. This way the research approach also has a strong epistemological component

interacting with the cognitive approach. The genetic decomposition then forms the basis of an instructional design that is implemented. After extensive observation and student interviews a renewed cycle is performed, which may cause changes in the decomposition and the design, and ultimately also in the theory. The nal genetic decomposition described consists of seven steps (see pp. 177178), which were materialised in the instructional design. Evidence for students constructions targeted in the different steps of the decomposition is provided by analyses of interview protocols. Some conclusions about concept development are made, indicating that a dynamic conception of limit is much more complicated than a process that is captured by the interiorization of an action (p. 190). A strong such conception is needed to move to a formal conception of limit, which is not static but instead is a very complex schema with important dynamic aspects and requires students to have constructed strong conceptions of quantication (p. 190). 3.2 Reasoning and beliefs In a study by Alcock and Simpson (2004, 2005), the interaction between students modes of reasoning (i.e. visual or non-visual) and their beliefs about their own role as learners is investigated. The research is a naturalistic inquiry into learners thinking about introductory real analysis (Alcock and Simpson 2004, p. 2), with the goal of the study being to develop a theory of the interactions between various aspects of students thinking (p. 7). The approach is thus cognitive and the research framework conceptual, since the study uses theoretical concepts from various sources rather than one overarching theory. Examples of such theoretical concepts used are on visualisation, concept image, spontaneous conceptions (Cornu 1991), perceptual proof scheme (Harel and Sowder 1998), semiotic control (Ferrari 2002), and, for the method, grounded theory, and the distinction account of/account for (Mason 2002). The empirical data consist of protocols from interviews with pairs of students, engaged in rst-year analysis courses, discussing general issues on university studies, working on given limit problems on sequences and series, and a review of the task session discussing proof and denitions. From the data the observed group of students could be classied either as visual or non-visual depending on their tendencies to introduce diagrams or not during tasks, to use gestures/qualitative terms or algebraic representations when offering explanations, to explicitly state their preference or disinclination for pictures or diagrams in reasoning, and to base their sense making to nonalgebraic or algebraic reasoning. The visualizers generally set focus on the mathematical objects as constructs, draw quick initial conclusions,

123

194

C. Bergsten

and show Conviction in their own assertions (Alcock and Simpson 2004, p. 10). However, a further analysis revealed three bands of behaviour of the visualizers, depending on the consistency of the way the mathematical objects were displayed with the formal denitions, and on the ability to use those denitions as a basis for argumentation. These behaviours were found to interact with the students beliefs about the learners role. Students that expect to see consistency and structure and use exible links between visual and formal representations in mathematics, show an internal sense of authority, setting value to their own judgement (p. 18). Students using images that are not of sufcient generality to justify their reasoning exhibit a belief that mathematics will be provided by an external authority (p. 24). In a similar way, the non-visual students could be divided into three bands of behaviour, depending on the accurate use of the mathematical denitions, and on the degree of semiotic control connecting algebraic representations with underlying concepts. Also the mathematical behaviour of these students revealed an interaction with their beliefs regarding internal or external authority. The way the course was conducted could not explain the different preferences, and both groups showed a wide range of success and failure, indicating that there is no perfect presentation that will be available to all students and successful (Alcock and Simpson 2005, p. 98). 3.3 The algebra and the topology of limits et al. (2005) is located The research presented in Barbe within the Anthropological Theory of Didactics (ATD) and adopts a theoretical framework where mathematical activity has to be interpreted as an ordinary human activity, along with other forms of activity, and thus proposes a general model of human activities (the praxeologies) that links and gives the same importance to their theoretical (knowledge) and practical (knowhow) dimension (Bosch and Gascon 2006, p. 59). A key object of study is therefore the organisation of the mathematical knowledge to be taught. Based on an epistemological approach, one of the main methodological principles of this research is to take into account how the mathematical knowledge as it is proposed to be taught constraints the students (and the teachers) mathematical practices. In the case of limits of functions, due to a complex historical process of didactic transposition, the mathematical knowledge to be taught appears to be a disconnected union of two mathematical organisations originated from different fundamental questions in the scholar mathematical institution: the algebra of limits that starts from the supposition of the existence of the limit of a function and poses the problem of how to calculate it

for a given family of functions; and the topology of limits approaching the question of the nature of the mathematical object limit of a function and responding to the problem of the existence of the limit of different kinds of functions. Due to traditional tasks and techniques in textbooks and syllabi, the algebra of limits becomes the practical dimension of the mathematical organisation to be taught, while at the same time the theoretical dimension remains closer to the topology of limits. This mismatch of the two parts of the taught praxeology causes problems for the teacher, as well as the students, to explain, justify, and give meaning to the work on limits. The available theoretical discourse is not appropriate to justify the techniques students learn to use and thus appears to be unmotivated, without any rationale and unable to justify the practice of the algebra of limitswhich, for this reason, tends to be considered as a mechanical practice difcult to develop. According to the ATD, the main reason for this phenomenon has to be found, not in the human cognition of teachers and students, but in the severe constraints imposed by the process of didactic transposition on the kind of mathematics that can be taught (and learned) at school. Without taking into account these institutional constraints, it seems difcult to understand what teachers and students do (and cannot do) when facing a problem involving limits of functions. The split mathematical praxeology about limits of functions explains some important distortions on the teachers and the students practice that are due to constraints coming from the rst steps of the process of didactic transposition. For instance, the difculties for the teacher to give meaning to the mathematical praxeologies to be taught, because the rationale of limits of functions (why we need to consider and calculate them) cannot be integrated in the mathematical practice that is actually developed at this level. The empirical data for analysing these issues in the particular case reported, were taken from syllabi, textbooks, and classroom observations. 3.4 An analysis of inuences of theory An overview of the inuence of the adopted theoretical basis on each of the three studies presented above is shown in Table 1, structured by the research question stated in Sect. 1, and by the descriptions, terms and models discussed above in the section on research frameworks and theories. For the main purpose categories, see Fig. 2. The approach has been classied to be cognitive, social, or epistemological, as described above, and the research framework theoretical, conceptual, or practical. The main research question of the studies, in my interpretation, are listed under questions. Then some key words describing research methods are given, with the kind of data used in

123

On the inuence of theory on research in mathematics education Table 1 A comparison of three research studies on limits of functions Study Main purpose Approach Research framework Questions Cottrill et al. (1996) Improved understanding and products Cognitive Theoretical: APOS theory How does a genetic decomposition of how the limit concept can be learned look like? Research cycle: analysisdesign implementationobservation analysis Interview protocols Dynamic conception of limit complicated Formal concept of limit not static Rened genetic decomposition of limit Further research on quantication needed, along with the genetic decomposition, to design effective instruction Alcock and Simpson (2004, 2005) Improved understanding Cognitive Conceptual: a set of local theories and concepts How do various aspects of students thinking interact? Open and task based interviews et al. (2005) Barbe Improved understanding Epistemological Theoretical: ATD

195

How are teachers practices restricted by mathematical and didactical phenomena? Epistemological analysis and observations of mathematical and didactical organisations Syllabi, textbooks, classroom observations The internal dynamic of the didactic process is affected by mathematical and didactical constraints that determine teachers practice and the mathematics taught Problems of motivation, meaning, atomisation of curricula, etc., need a deeper understanding of institutional restrictions regulating teaching

Methods

Evidence Conclusions

Interview protocols A theory about the interactions between students tendency to visualize and beliefs about their own role as learners At least in small group teaching situations, different students tendencies to visualize should be taken into account

Implications

the study listed under evidence, and my interpretation of the main results under conclusions. Finally, some implications of the studies are added, relating to further research needed, educational practice or recommendations. The two studies using a cognitive approach both investigate the inuence of learning environments on the development of students understanding of the mathematical concept of limit. The research frameworks adopted, however, may be characterized as closed and open, respectively. Cottrill et al. (1996) start with, and stay within, a specic theory focusing, along with an epistemological analysis of the limit concept, on the cognitive development of the individual student. This stance is forcing interview data to be interpreted in terms of the basic notions of the theory only, that is actions, processes, objects, and schemas: In trying to t our observations with the APOS theory, we felt the need to pay more attention to the idea of schema than in our previous work with this theory (p. 190). The clinical interview is chosen, in line with the Piaget tradition, as the method for collecting evidence on the state of a students mental schema. Due to the double aim of the study, a cyclic process of research is needed to rene the genetic decomposition. This is a closed (theoretical) framework, and the conclusions may be called a progressive conrmation of the expected aims. As a contrast, the study by Alcock and Simpson (2004) began as a qualitative investigation of the way different

learning environments inuence students developing understanding of real analysis (p. 1). The centrality of the distinction between visualizers and non-visualizers, and the interacting role of beliefs, did only emerge by inductive analysis of the data (p. 1). This is an indicator of a kind of openness of the adopted conceptual framework. Here, the aim was not to develop an instructional design by using a specic theory-based tool, but to increase understanding of the inuence of learning environments on students conceptual understanding. Thus, possibly not to force students thinking to t a specic line of development, the data collection method chosen was task solving in pairs, in addition to open questions on general views on mathematics and on proof and denitions. Based on the conceptual framework, which can be seen as emerging from the research problem and the interpretation of data, the conclusion of the research is the development of a theory which accounts for the students behaviour based on the interactions between degrees of visualization and beliefs on authority (Alcock and Simpson 2004, p. 2). et al. (2005) shares with Cottrill The study by Barbe et al. (1996) a questioning of the mathematical content in use but outlines a very different kind of questioning of this object. While Alcock and Simpson (2004, 2005) take the scholar point of view on limits of functions for granted, the theory of didactic transposition allows this questioning. By the analytical tools provided by the ATD

123

196

C. Bergsten

framework the problems that teachers and students encounter are found to be located in the disconnectedness of the mathematical and didactical organisation of the knowledge, due to the process of didactic transposition and institutional constraints. Focus is more on the didactical organisation of teaching and related phenomena than on individual students thinking processes. Taking this epistemological approach, the empirical investigations reveal at a systemic level the potential meaning and motivation of the students work by looking at what types of problems and techniques are focused in the classroom, and how these are related to an available theoretical discourse that justies their use. The research question, methodological tools, the theoretical discourse for analysing the data, as well as the formulation of the results and conclusions, are all expressed in terms of the predened theoretical framework. The three studies highlighted in the analysis all originate from common observations of student difculties in the mathematical content area of limits of functions, but display, by their different approaches and research frameworks, different kinds and levels of research questions and answers, based on different kinds of methods and evidence. The conclusions from the research, in particular and as a consequence of the research questions, differ considerably at a qualitative level: within the APOS theory, claims are made at a local conceptual and instructional level; within the conceptual framework, a local theory to account for the data is postulated, with a focus on the individual students thinking; and within the ATD framework, explanations are found at a systemic level. In addition, the implications listed in Table 1 stay for the cognitive approach at a local level of understandings and instruction, while the epistemological approach takes another perspective and considers the level of institutional restrictions as necessary to account for teachers practice and students behaviour.

4.1 The inuence of theory on the research Of the ways described by Silver and Herbst (2007) in which theory can be involved as a mediator between research and problems (p. 50), Table 1 shows how at least three of them came into play in the comparison of studies discussed here, that is to establish a research frame to a problem, transform a problem to research questions, and as generator of problems for research. The different approaches transformed the commonsensical problem of observed difculties when teaching and learning limits of functions into a research problem formulated within a conceptual framework or different theoretical frameworks. As a mediator between research and practice, theory may according to Silver and Herbst (2007, p. 56) serve the role of supporting prescription (setting up an ideal model for evaluating practice), understanding (deciding on the best practice as derived from empirical research on existing practices), and prediction (providing a denition of a best practice). In the case of Cottrill et al. (1996), the cyclic research and instructional design process activates all these roles of the APOS theory. The study by Alcock and Simpson (2004, 2005) was not focused on instructional practice, and the recommendations for aspects of practice provided (see the category implications in Table 1) have no support in the research to say anything about how students tendencies to visualize should be taken into account in teaching situations. The category of serving to understand practice is evident in et al. (2005), but as argued by Silver and Herbst Barbe (2007), even the most descriptive approaches of research to practice include a prescription of what that practice should be that allows it to be visible and isolated from the rest of experience (p. 53). This is a kind of methodological aspect of prescription present in any research using theoretical tools to dene and categorise its objects of study. According to Silver and Herbst (2007), an analysis of the role of theory as mediator between problems and practice requires a discussion of the complex terms problem and practice. However, it mainly refers to the identication of what in a practice is seen as a problem, or from what practice or aspects of a practice a problem originates. It may also concern how new practices can be developed from the study of problems. All these roles of theory are present in et al. (2005) and central to the approach. In the case Barbe of Alcock and Simpson (2004, 2005), practice is again not the focus but an understanding of interactions between different aspects of students thinking. Neither the conceptual framework nor the resulting local theory plays any clear mediating roles between problems and practice. The research process in Cottrill et al. (1996) dened a chain of developing practices where the theory used helped to identify problems within the practice, establish criteria for a successful development, and help redesign the practice.

4 Discussion To compare research studies is in itself a research task, and, as a consequence, requires a theoretical stance within which to work. Within the scholarship triangle (Fig. 1) it concerns theory as a mediator between research and problems, a role described by Silver and Herbst (2007, p. 50) as theory as systematic organization of a corpus of research on a problem. In this paper analytical tools based mainly on Cobb (2007), Lester (2005), Niss (2007a, b), and Silver and Herbst (2007) were chosen to organise the present meta study by providing a basis for the structure of Table 1 and the discussion of the ndings.

123

On the inuence of theory on research in mathematics education

197

This analysis of the mediating roles of theory between problems, practice, and research, based on the scholarship triangle (Fig. 1), shows how the two research studies with a theoretical framework included more extensive and elaborated use of theory than the study structured by a conceptual framework. It can also be noted that for the two theoretical framework studies, theory development is conned to the internal level, that is the theoretical framework itself. In contrast, the theoretical outcome of the conceptual framework study is external, in the emergence of a new (local) theory. The criteria used in Cobb (2007) for comparing theoretical perspectives, that is characterization of the individual, usefulness for instructional design, and limitations, may also be relevant for comparing individual studies with a focus on their theoretical underpinnings. His summary description for how these criteria characterize cognitive psychology ts well to the outcome of the comparison made here of the cognitive approach studies, for example seeing the epistemic individual as reorganiser of activity (p. 28). However, taking an epistemological approach with a focus on the structure of mathematical knowledge and its organisation in didactical settings, the criteria on the characterization of the individual is not relevant. Concerning the pragmatic criterion of usefulness for instructional design, the cognitive approach studies help the design of educational activities but the means for supporting learning are conned to the level of tasks (cf. Cobb 2007, p. 28), while the outcome of the study using the epistemological approach is useful at a curriculum level and for modelling the didactical organisation of knowledge in the classroom. The common basis for the three studies is the commonsensical problem that limit of functions is a difcult topic for students. Adopting a specic approach inuences what kinds of questions can be asked about this problem, with a purpose to improve understanding or to improve products (instruction and learning). The research framework denes the discourse within which the research questions are formulated, and their answers. From the research reports it is impossible to trace the reasons for these choices, other than the general orientations and interests of the researchers. According to Cobb (2007, p. 9), considering mathematics education as a design science, the choice of theoretical perspective requires pragmatic justication. This comment leads the discussion to the issue of interpreting research results. 4.2 Comparing outcomes of research It is evident from the examples in this meta study how a chosen research framework denes the world in which the research lives and grows, a fact that also has implications on

how to interpret outcomes of research. It points to the difcult task to compare research results within a common eld of study taking into account the different approaches and research frameworks used. For example, the fact that institutional constraints rarely are taken into account in didactic research makes it difcult to compare results from studies doing that with others not focusing on those constraints. In Bosch, Chevallard and Gascon (2006) such a comparison between two studies on the concept of continuity is found, focusing on consequences of considering several dimensions of a mathematical practice instead of only one, concluding that students difculties in the learning of a piece of knowledge that is praxeologically out of meaning can be taken as a positive symptom of the educational system, instead of a problem in itself (p. 1261). It is in the nature of a scientic approach to adopt a specic focus in order to reveal an observation of any signicance. Paradoxical as it may seem, it is necessary to restrict attention in order to account for the inherent complexity of the studied phenomenon, to avoid simplications. This implies however, that after doing such negrained analyses, which may be situated at different locations on a micromacro level scale in terms of individual cognitive functioning or institutional framings (see the paper by Prediger in this issue), it is also necessary to put the resulting observations into context to validate any further conclusions or implications. For example, how does the observed covariation between students tendency to visualize and beliefs about their own role as learners in the study by Alcock and Simpson (2004, 2005) link to the construction of the learners cognitive schemas of a specic mathematical object like limits of functions, as developed through the seven steps of the genetic decomposition outlined in Cottrill et al. (1996)? And how does it relate to the connectedness of the mathematical organisation as dened by the curriculum? The different studies focus on different dimensions and offer solutions to different kinds of research problems, which may be related to a recognised general (commonsensical) problem of the difculties of teaching and learning the mathematical content area limits of functions. A way to understand the nature of such relations may be by pragmatic justication. An important difference found between the studies analysed here is the level of intervention they suggest in their conclusions and implications, as indicated in Table 1. Alcock and Simpson (2004) propose modifying teaching by taking into account student characteristics and beliefs, possible only at an individual or small group level. To accomplish this they recommend work with specic kinds of tasks (cf. Cobb, 2007). However, the approach and research framework adopted do not provide much advice for how to design such tasks or how to work with them. Cottrill et al. (1996) take the design and effectiveness of

123

198

C. Bergsten

the teaching proposal into account without questioning if limits of functions have to be taught or why, as well as the tre) of this specic content connections (or the raison de within the mathematics curriculum of which it is a part. et al. (2005) point out the main reason of the difBarbe culty of learning limits at an institutional level. Students have difculties in learning, and teachers in teaching because the mathematical organisation does not have any sense in the practical mathematical activity carried out in school at this level. In each case very different objects are questioned: students beliefs and abilities, teaching designs, and the rationale of a given notion in a curriculum, respectively. The level of intervention thus varies completely. Allowing an oversimplication, one may formulate the observations above as no matter how good are the stu et al. dents or the teacher, it has no sense (for Barbe 2005), no matter how are the students if the teaching process is not good (for Cottrill et al. 1996), or no matter why this is taught nor how, what matters rst is the nature of the students (for Alcock and Simpson 2004, 2005). To synthesise such seemingly confusing results into a coherent knowledge base to support instructional design dealing with the particular commonsensical educational problem is an important but difcult research problem. Artigue et al. (2007) point to some examples where efforts have been made to combine different approaches to research on post-secondary level mathematics, stating that there is an explosion of notions and terms that is not easy to make sense of, but links and partial translations are often possible (p. 1030). However, by adopting different approaches and theoretical frameworks, the discussion above has shown that the objects of study also differ and cannot be directly translated into each other. This observation is a challenge to the development of networking strategies.

is the appearance of the second wave theories in cognitive psychology, due to the need to account also for affective and contextual factors in cognition (De Corte et al. 1996). Didactical research therefore needs to investigate all these dimensions with an awareness that there is one more step to go to obtain a progression of our didactical knowledge of a content area like limits of functions. This step requires a development of a (hierarchically organised) network of the different contributions to this knowledge. Considering the implications of constraints of different research frameworks, examples of which have been demonstrated above, ways to link research results on different dimensions of didactical phenomena, such as epistemological, cognitive, and social, need to be developed. Comparisons of the kind presented here, considering the assumptions and characteristics of the research process on which the results and their implications depend, may contribute to a basis for the construction of such networks of didactical knowledge. References
Alcock, L., & Simpson, A. (2004). Convergence of sequences and series: Interactions between visual reasoning and the learners beliefs about their own role. Educational Studies in Mathematics, 57, 132. Alcock, L., & Simpson, A. (2005). Convergence of sequences and series 2: Interactions between nonvisual reasoning and the learners beliefs about their own role. Educational Studies in Mathematics, 58, 77100. Artigue, M., Batanero, C., & Kent, P. (2007). Mathematics thinking and learning at post-secondary level. In F. K. Lester (Ed.), Second handbook of research on mathematics teaching and learning (pp. 10111049). Charlotte, NC: Information Age Publishing. , J., Bosch, M., Espinoza, L., & Gascon, J. (2005). Didactic Barbe restrictions on the teachers practice: The case of limits of functions in Spanish high schools. Educational Studies in Mathematics, 59, 235268. Bergsten, C. (2006). Trying to reach the limitThe role of algebra in mathematical reasoning. In J. Novotna et al. (Eds.), Mathematics in the centre. Proceedings of PME 30 (Vol. 2, pp. 153160). Prague: Charles University. Bikner-Ahsbahs, A., & Prediger, S. (2006). Diversity of theories in mathematics educationHow can we deal with it? Zentralblatt fur Didaktik der Mathematik, 38, 5257. Bosch, M., Chevallard, Y., & Gascon, J. (2006). Science or magic? The use of models and theories in didactics of mathematics. In M. Bosch (Eds.), European research in mathematics education IV, Proceedings of CERME 4 (pp. 12541263). Barcelona. Bosch, M., & Gascon, J. (2006). Twenty-ve years of the didactic transposition. Bulletin of the International Commission on Mathematical Instruction, No, 58, 5165. Burn, B. (2005). The vice: Some historically inspired and proofgenerated steps to limits of sequences. Educational Studies in Mathematics, 60, 269295. Cobb, P. (2007). Putting philosophy to work. Coping with multiple research perspectives. In F. K. Lester (Ed.), Second handbook of research on mathematics teaching and learning (pp. 337). Charlotte, NC: Information Age Publishing.

5 Conclusion The comparative meta study made here shows, while seeing the content eld of the didactics of limits of function as a whole from both epistemic and pragmatic purposes, that research results from one approach only cannot account for all problematic phenomena, including their causes and potential solutions. Using a specic research framework connes the world in which the research and its implications live, and there are many worlds involved in the context of educational systems, such as the cognitive, the social and the epistemological. However, these worlds are analytic categories inserted by our ways of organising what we experience as reality, and are thus not separated in the continuous ow of life. An evidence of this

123

On the inuence of theory on research in mathematics education Cornu, B. (1991). Limits. In D. Tall (Ed.), Advanced mathematical thinking (pp. 153166). Dordrecht: Kluwer. Cottrill, J., Dubinsky, E., Nichols, D., Schwingendorf, K., Thomas, K., & Vidakovic, D. (1996). Understanding the limit concept: Beginning with a coordinated process scheme. Journal of Mathematical Behavior, 15, 167192. De Corte, E., Greer, B., & Verschaffel, L. (1996). Mathematics learning and teaching. In D. Berliner, & R. Galfee (Eds.), Handbook of educational psychology (pp. 491549). New York: Macmillan. Eisenhart, M. (1991). Conceptual frameworks for research circa 1991: Ideas from a cultural anthropologist; implications for mathematics education researchers. In Proceedings of the 13th annual meeting of the North American Chapter of PME (Vol. 1, pp. 202219). Blacksburg, VA. Ferrari, P.L. (2002). Understanding elementary number theory at the undergraduate level: A semiotic approach. In S. R. Campbell, & R. Zazkis (Eds.), Learning and teaching number theory: Research in cognition and instruction (pp. 97115). Westport, CT: Ablex Publishing. Gravemeijer, K. (1994). Educational development and developmental research in mathematics education. Journal for Research in Mathematics Education, 25, 443471. Harel, G., & Sowder, L. (1998). Students proof schemes: results from explorative studies. In A. Shoenfeld, J. Kaput, & E. Dubinsky (Eds.), CBMS issues in mathematics education III, pp. 234283. Harel, G., & Trgalova, J. (1996). Higher mathematics education. In A. Bishop, K. Clements, C. Keitel, J. Kilpatrick, & C. Laborde (Eds.), International handbook of mathematics education. (Part 1, pp. 675700). Dordrecht: Kluwer. Juter, K. (2006). Limits of functionsUniversity students concept University of development. Doctoral Thesis 2006:08. Lulea Technology, Sweden. Lester, F. (2005). On the theoretical, conceptual, and philosophical foundations for research in mathematics education. Zentralblatt fu r Didaktik der Mathematik, 37, 457467. Mamona-Downs, J. (2001). Letting the intuitive bear on the formal; a didactical approach for the understanding of the limit of a sequence. Educational Studies in Mathematics, 48, 259288.

199 Mason, J. (2002). Researching your own practice: The discipline of noticing. London: Routledge Farmer. Mason, J., & Waywood, A. (1996). The role of theory in mathematics education research. In A. Bishop, K. Clements, C. Keitel, J. Kilpatrick, & C. Laborde (Eds.), International handbook of mathematics education (pp. 10551089). Dordrecht: Kluwer. Niss, M. (1999). Aspects of the nature and state of research in mathematics education. Educational Studies in Mathematics, 40, 124. Niss, M. (2007a). The concept and role of theory in mathematics sval, & F. education. In C. Bergsten, B. Grevholm, H. Ma Rnning (Eds.), Relating practice and research in mathematics education. Proceedings of Norma 05 (pp. 97110) Trondheim: Tapir. Niss, M. (2007b). Reections on the state of and trends in research on mathematics teaching, learning. From here to Utopia. In F. K. Lester (Ed.), Second handbook of research on mathematics teaching and learning (pp. 12931312). Charlotte, NC: Information Age Publishing. Przenioslo, J. (2005). Introducing the concept of convergence of a sequence in secondary school. Educational Studies in Mathematics, 60, 7193. Sierpinska, A., & Lerman, S. (1996). Epistemologies of mathematics and of mathematics education. In A. Bishop, K. Clements, C. Keitel, J. Kilpatrick, & C. Laborde (Eds.), International handbook of mathematics education (pp. 827876). Dordrecht: Kluwer. Silver, E., & Herbst, P. (2007). Theory in mathematics education scholarship. In F. K. Lester (Ed.), Second handbook of research on mathematics teaching and learning (pp. 3967). Charlotte, NC: Information Age Publishing. Sriraman, B, & English, L. (2005). Theories of mathematics education: A global survey of theoretical FRAMEWORKS/ trends in mathematics education research. Zentralblatt fu r Didaktik der Mathematik, 37, 450456. Tall, D. (2004). Thinking through three worlds of mathematics. In M. Johnsen Hines, & A.B. Fuglestad (Eds.), Proceedings of PME 28 (pp. 281288). Norway: Bergen University College.

123