Warning Concerning Copyright Restrictions The Copyright law of the United States (Title 17, United States Code) governs the making of photocopies or other reproductions of copyright material. Under certain conditions specified in the law, libraries and archives are authorized to furnish a photocopy or other reproduction. One of these specified conditions is that the photocopy or reproduction not be “used for any purposes other than private study, scholarship, or research.” If a user makes a request for, or later uses, a photocopy or reproduction for purposes in excess of “fair use,” that user may be liable for copyright infringement.

-- Computers, Cognition, and Work

A series edited by Gary M. Olson, Judith S. Olson, and Bill Curtis

! CSCL: THEORY AND PRACTICE Pox. The Human Tutorial Dialogue Project Issues in the Design of lnstructiona( I OF AN EMERGING PARADIGM Systems

Hoschka (Ed.). Computers as Assistants: A New Generation of Support Systems Koschmann (Ed.) l CSCL: Theory and Practice of an Emerging Paradigm Moran/ Carroll (Eds.) l Design Rationale: Concepts, Techniques, and Use Oppermann (Ed.) l Adaptive User Support: Ergonomic Design of Manually and

Automatically Adaptable Software Smith . Collective Intelligence in Computer- Based Collaboration

lElA I996 Edited by

Timothy Koschmann Southern Illinois University

LAWRENCE ERLBAUM ASSOCIATES, PUBLISHERS Mahwah, New Jersey

CONTRtBUTORS X

m& i Mo~‘ riSTison BBN CorpWatioo

ch, istitte hf. Neuwirth Dept. of English Carnegie Mellon University

,&@ met Riel hltdearn

kremy Roschelfe Institute for Research on Learning

Marlene Scardamalio Dotado fostitute for Studies in Education of the University of Toronto

Elliot sdwoy Dept. of EECS University of Michigan

Rand J. Spiro center for the Study of Reading University of Illinois, Urbana- Champaign

Patricia G. W’qiohn Dept. of English Carnegie Mellon University

PREFACE This is a book about a newly emerging area of research in instructional technology. Because it is changing so rapidly, it is difficult to say much about it that is not, or will not in the near future become, subject to contradiction. We use the acronym “CSCL” in this volume to refer to this emerging area of work, though even this title is a source of controversy. Initially, CSCL was* chosen as an acronym for Computer- Supported Collaborative Learning, but as you will see there are some who are uncomfortable with this label. For example, in his chapter in this volume, Roy Pea argues that “collaborative” h ? is often not a descriptive term for what learners do in instructional settings. Further, as the field develops, the technology that is used to support col- laboration may not always involve computers, at least not in the direct ways that computers have been used to support instruction in the past. Interest is growing, for example, in the use of video conferencing and other commu- nications technologies to support collaborative methods of instruction. To avoid getting bogged down in this terminological debate, we simply use ‘CSCL” in this book as a designation in its own right, making no claims about what it might actually stand for.

The inspiration for this volume was a symposium that took place at the 1992 annual meeting of the American Educational Research Association (AERA). The symposium, entitled “Instructional Theories Underlying the Use of Networked Computers in the Classroom,” was organized to give several prominent CSCL developers an opportunity to articulate the theories of learning and instruction upon which their work was founded. Papers result-

xi

xii PREFACE lng from that symposium later appeared in a special issue of the Journal of kami, rg s& rces. Earlier versions of three of the chapters in this book (chapters 4, 7, and 10) first appeared in that issue. This volume represents an attempt to broaden the discussion begun in the AERA symposium by Inviting additional researchers to present their perspectives.

The authors of these chapters talk a lot about the theory underlying their work. In part, this is because that is what they were asked to do, but beyond this ] think it is also an indication of the state of the field. In an established paradigm in which the theories and methods are well agreed on. such discussion becomes unnecessary. CSCL, however, has not yet reached the stage of “normal” science. There is much to be worked out yet, and we hope that this book will play some role in helping to define a direction for future work in this field.

The chapters appear (excepting the introductory chapter and the after- word by Janet Kolodner and Mark Guzdial) in alphabetical order based on the first authors’ surnames. This was done not for lack of a better way of organizing the chapters, but because the organizational possibilities were too numerous and I did not wish to privilege one over another. The chapters, for example, could have been divided based on various aspects of the applications described (e. g., planned locus of use, instructional role, how use is coordinated in time), the domain of instruction (e. g., physics, medi- cine, English composition), the level of instruction (i. e., elementary, secon- dary, college, professional), or by the theoretical dispositions of the system designers. Instead, by not imposing a structure on this collection, I hoped that you the reader might feel freer to explore the chapters in a way that best suits your needs. You will be aided in this exploration by the inclusion of comprehensive author and subject indices at the back of the book

ACKNOWLEDGMENTS I would like to acknowledge the numerous sources of help I have had in bringing this collection together. Hollis Heimbouch, my editor at Lawrence Erlbaum until her departure near the completion of the project, was a continuous source of support and encouragement. Nancy Sanchez provided extensive help with the indexing and other logistical details. Finally, while editing this volume, I was partially supported by a Spencer Post- Graduate Fellowship from the National Academy of Education.

The quote from Thomas Kuhn’s The Structure of Scientific Revdutions on the following page appears with permission of the University of Chicago Press. Earlier versions of chapters 4, 7.9, and 10 appeared in The Journal of the Learning Sciences, published by Lawrence Erlbaum Associates.

These examples point to the. rnmt fundamental aspect dthe incommensurab& ty d competing paradigw. In a sense that I am unable to explicate further, the proponents of competing paradigms practice their trades in difierent worlds. Practicing in different worlds. the two groups af scientists see diNerent things when they look from the same point in the same direction. Again, that is not to say that they can see anythirq they please. Bdh are Icoking at the world and what they laok ot has not chang& But in some areas they see different things, and they see them in different relations one lo the other. That is why. &fare they can hape to communicate futtx on? gmup or the other must experience the conversion that we haoe been calIing a par&& m shiA: a

-Thomas S. Kuhn. The Sbucture OrScientific Revdutions (p. 150) *. 3

CHAPTER 1

PARADIGM SHIFTS AND INSTRUCTIONAL TECHNOLOGY:

AN INTRODUCTION Timothy Koschmann Southern Illinois University

In his well- known essay on the nature of scientific revolutions, Kuhn (1972) theorized that scientific research proceeds through long, relatively stable periods of normal science intermittently punctuated by briefer, more tumul- tuous times in which new paradigms for research may emerge. He charac- terized normal science as “research firmly based upon one or more past scientific achievements, achievements that some particular scientific commu- nity acknowledges for a time as supplying the foundation for its further practice” (p. 10).

A scientiftc achievement represents a paradigm for Kuhn if it raises a compelling set of researchable questions and attracts a following of workers 5. intent on pursuing those questions. The paradigm supplies its practitioners with “topics, tools, methodologies, and premises” (Lehnert, 1984, p. 22). It provides purchase in attacking what might previously have been considered intractable problems. A paradigm is not fixed, however, but is refined and extended through use. In Kuhn’s words, it becomes “an object for further articulation and specification under new and stringent conditions” (1972, p. 23). Over time, competing paradigms may emerge, potentially leading to one paradigm’s abandonment in favor of another. Such shifts are always revo- lutionary occurrences. As Kuhn observed, “the transition between compet- ing paradigms cannot be made a step at a time, forced by logic and neutral experience. Like the gestalt switch, it must occur all at once (though not necessarily in an instant) or not at all” (1972, p. 150).

One interesting feature of Kuhn’s theory of scientific revolutions is what he referred to as the “incommensurability of the pre and post- revolutionary

--- 2 KOSCHMANN normal- scientific traditions” (1972, p. 148). Adherents to a new paradigm adopt an altered Weltunanschauung, prescribing a new way of observing, reflecting on, and describing the world. Though the notion of incommen- surability is a source of controversy among philosophers of science (Biagioii, 1990; Khcher, l978), Kuhn held that the effect of a paradigm shift is to produce a divided community of researchers no longer able to debate their respective posltions, owing to fundamental differences in terminology, conceptual frame- works, and views on what constitutes the legitimate questions of science.

In this chapter i argue that, seen from a Kuhnian perspective, instruc- tional technology (IT) has undergone several such paradigmatic shifts in its relatively brief history? As a result of these shifts, the field has been balkan- axed Into a number of smaller communities, each utilizing different research l, m& ee and espousing largely Incommensurable views of learning and lnstr~& I. I argue further that there now appears to be a new paradigm merging within IT, arising from yet another perspective on these same Issues. This developing paradigm, for which the acronym CSCL has been coined (Koschmann, 1994a), focuses on the use of technology as a medfa- tlonal tool within collaborative methods of instruction. Before pursing this ‘analysis, however, let me address some potential concerns about the iegiti- macy of applying Kuhn’s theories to the body of work devoted to the uses of technology in instruction.

First in this regard is the issue of natural versus artificial science. In Sciences of rhe Artificial, Simon (1969) defined natural science as “a body of knowledge about some class of things- objects or phenomena- in the world; about the characteristics and properties that they have: about how they behave and interact with each other” (p. 1). The historical events on which Kuhn focused, such as Lavoisier’s discovery of oxygen and Copernicus’ development of a new model of the solar system, were clearly examples of this type of endeavor. The central thrust of work in IT, on the other hand, has been to produce practical artifacts to support instruction rather than to discover new principles about the natural world. Simon proposed an alternative category of scientific inquiry (i. e., artificial science) for work in areas devoted to the production of teleological objects designed to serve a particular goal or purpose. The issue, therefore, is whether or not it is appropriate to generalize Kuhn’s descriptions of conduct within the natural sciences to work within an artificial science, such as IT.

A second, and related, concern has to do with the role of theory in the emergence and dissolution of research paradigms. Thagard (1992) has argued that although there have been noteworthy conceptual shifts in the social sciences, such as the shift in psychology from behaviorism to more cognitive approaches, they are different from the revolutionary shifts that have oc- curred in the natural sciences. He made acritical distinction between theories

$ll?&?+ PP roaches. Thagard defined a theory as a “coherent collection of 1. PARADIGM SHIFTS AND INSTRUCTIONAL TECHNOLOGY 3

hypotheses, [which] serve to explain a broad range of empirical generafiza- tions and facts” and an approach as “a general collection of experimental methods and explanatory styles” (1992. p. 225). He concluded that because the social sciences have. falied to produce any broad, unifying theorles~ comparable to Newton’s theories of mechanics or Darwin’s theory of natural selection, the conceptual shifts that have marked past research in these fields were Ymore the result of methodological considerations than evaluations of explanatory coherence” (p. 225). Thagard’s position is of interest here be- cause I argue that the shifts that have occurred in IT were in fact driven by shifts in underlying psychological theories of learning and instruction.

Whereas it is quite true that instructional technology, as a field of study, is different in many respects from the scientific disciplines described by Kuhn, this does not mean that it could not be productively studied by the same means. Although the practices of research and standards of evidence utilized within a field such as IT may be quite different from those employed within the natural sciences, there is no reason to believe that the cultural factors that organize and lend structure to the field would be any different from the analogous factors operating within the disciplines studied by Kuhn. By the same token, Thagard’s distinction between theories and approaches, al- though important to his typoiogy of conceptual shifts, does not preclude an historical analysis of work within IT. Although the underlying theories of learning and instruction that I argue have informed work in IT do not meet Thagard’s standard for a “theory,” the fact that they have resulted in paradig- matic shifts in practice is the important issue here. Whether we choose to call the fundamental reconceptualizations underlying these shifts “changes in theory” or “changes in approach” is of little consequence to this discussion.

Conducting a Kuhnian analysis of IT is an instructive exercise, requiring a reexamination of the theories that have motivated work in the field and the practices by which technological innovations are designed and evalu- ated. Focusing on foundational theories and research practices, as opposed to the form and intended role of the designed artifacts, represents a novel way of conceptualizing past (and future) work. I begin this analysis by looking briefly at some of the past paradigms for research in the field. This serves as background to the more central question of this chapter; that is, does the emerging body of work devoted to CSCL constitute a new paradigm for research in IT7

PAST PARADIGMS OF INSTRUCTIONAL TECHNOLOGY

There are many ways of using technology to support instruction. Before computers, a number of other forms of technology- film, radio, and televl- sion- had been introduced into the classroom with varying degrees of sue- cess (Cuban, 19%). It was not until the advent of computers, however, that

4 KOSCHMANN instructional technology came into its own as a broad area of study and my analysis, therefore, focuses on the use of computer- based technologies. ’ One ca identify several past paradigms for the instructional use of technology, both within and outside of the classroom. In this section, 1 describe three- Comnuter- Assisted Instruction (CAI), Intelligent Tutoring Systems (ITSs), and -~ . the Logo- as- Latin Paradigm.

Because the paradigms we are about to consider are paradigms in edu- cational technology, I endeavor to address four questions for each- two theory- based, and two relating to practice. First, what is the implicit theory offeoming upon which the paradigm was constructed? Formulating an an- swer to this question will in many cases entail an exploration of the para- digm’s epistemological commitments and its underlying philosophy of mind (Ernest, 1995). Second, what is the theory dpedagogv, that is, the underlying model of instruction implicit to the paradigm? Of particular interest here, of course, is the role of technology within this model. Shifting to the practical aspects of the paradigm, the third question explores its research methoddogy (i. e., How are claims warranted? What counts as scientific evidence? What are the methods by which this evidence is gathered?). The fourth and final question concerns what Kuhn called the “legitimate” (1972, p. 10) research nroblems of the paradigm, that is. what are the important research questions that the paradigm was established to address?

Developing an historical analysis of past paradigms for research in IT is an ambitious project to which a full book could be devoted. Because the focus of this volume is on the development of CSCL as an emerging area of work, 1 only provide a cursory sketch of the paradigms that have come before. 2 An exploration of this background material is essential, however, to developing an understanding of the context within which work in CSCL arises.

CAI Paradigm Because the term Computer- AssistedlnstrUcfion (CA]), along with related terms such a Computer- Based Instruction and Computer- Aided Learning, is used in

‘The term computer should be construed broadly enough. however, to include emerging technologies such as high- bandwidth networks. wireless telecommunications, interactive television. and video conferencing.

2For the reader interested in exploring this body of work in greater detail. there are a number of references that could serve as points of departure. O’Shea and Self (1983) provided an ew4fent ovwvtew of early work done within the CAI tradition. Larkin and Chabay (1992) highlighted some of the connections among more recent work in CAI and ongoirlg work within the ITS tradition. Wenger (1987) provides a very thoughtful analysis of work within the ITS tradition. The contrast between constructivist theory and mare traditional approaches to tnstructionaf design are taken up in a book edited by Duffy and Jonassen (1992). Finally, three edited collections( Jones & Winne, 1992; LaJoie & Derry. 1993; Rutkowska &Crook. 1987) straddle the dfvisfon between constructivist theories of education and traditional ITS research.

,. PARADIGM SHIFTS AND INSTRUCTIONAL TECHNOLOGY 5 a variety of ways in the IT literature, some clarification is required. In the early literature, CA was used generically as a blanket term for all uses of computers in education (e. g.. Steinberg, 1991). Later, it came to represent a default background against which other more specific approaches were contrasted (e. g., Wenger. 1987). In the current discussion, however, I use the term in a more specific sense to refer to a particular paradigm in the design and evaluation of instructional technologies. I have chosen IBM’s release of Coursewriter I, the first CAl authoring tool (Suppes & Macken, 1978) in 1960 to serve as the inaugural event for the emergence of this paradigm. 3 The advent of courseware building tools made it possible for individuals without formal training in programming or computer science to develop their own computer- based teaching aids. Because many CAl devel- opers have backgrounds in teaching (Larkin & Chabay, 1992) applications developed under this paradigm tend to be straightforward and practical instructional tools designed around the identified needs of the classroom. ’

Because of these close ties between CAI developers and education practi- tioners, CAI applications tend to reflect the beliefs and attitudes of the general education community. Cuban (1993) described what he referred to as the “dominant cultural norms” with respect to learning, instruction, and the nature of knowledge. These beliefs, though rarely made explicit, are pervasive within the education world and are embraced by students, teachers, school administrators, and members of the surrounding community.‘ fn this vtew, laamfnglsseen as thepasslveacquisltionor absorptionof anestabllsh@ d( and’.* often rigidly defined) body of information. The teacher’s role Is to “acquire formal knowledge, flnd efflclent ways of sharing it, and determine whether pupils have learned what was taught” (Cuban, 1993, p. 248). Instruction, then, becomes a process of transmission or delivery. Reflecting the influence of prior work in programmed instruction (Skinner, 1968) and instructional design( Gagn6,1968), CAfapplicationsutilizeastrategyofidentifyhrg aspecific set of learning goals, decomposing these goals into a set of simpler component

“In providing an historical account of past work in IT. I have identified specific events to mark the emergence of each of the paradigms described. By coincidence, each 01 theseselected events occurred at or near the beginnlng of a new decade. Thls pattern was quite accidental. however. and not meant to Imply that a shift in paradigms need be expected every ten years. Indeed, each selection was somewhat arbitrary and for every chosen event there were alternatives. before and after, that could have served in its place. Selecting alternative events would not only change the dates on which some of the shifts occurred, but could in some cases change the order of their emergence. This type of historical gerrymandering, however, would fn “0 WlY after the central CMIII of the chapter. namely that shifts in research practice how occurred in instructional technology, resulting in the creation of several distinct communftfes of practice.

‘At least this has been the intent. Cuban (1986) has argued that the failure of various technology- driven initiatives to achieve an appreciable impact has been due lar~ elv to a failure on the part of the designers to fully appreciate the expectations and requirements of classroom practitfoners.

KOSCHMANN 6

tasb, m~, finally, && oping a sequence of activities designed to eVentUallY lead to the &ievement of the Original learning objectives.

Bvaluative research fn education has been, and to a large extent contin- ues to be, dominated by a tradition that is both behavioristic and wperi- mentahst (Lagemmn, 1989). Work in CA1 can be seen as upholding this tradition (Blafsdell, 1976). Sharing the positivist’s distrust of nonpublic, men- tahstlc phenomena CAf researchers construe learning as a measurable difference tn displayed proficiency. Learning, so defined, serves as a depend- ent varfabfe fn CA1 research while the introduction of some form of techno- logical innovation represents the experimental intervention. The use of control conditions is common in CAl studies- either through actual matched samples or through the use of pre- and post- treatment testing in which experimental subjects serve as their own control.

CAI studies are designed to address the question: What are the instruc- tional benefits of an introduced technology? Research under this paradigm, therefore, has had as a central concern the issue of insfnrctiono/ eficucy. The paradigm itself has undergone some refinement over the years. Early work related to programmed instruction focused on parameters of reinforcement and their effects on learning (e. g., Coulsen. Estavan, Melaragno, and Silber- man, 1982; Gilman, 1967). These were carefully controlled laboratory studies very much in the style of the behavioristic school (Skinner, 1968). Later work (e. g., Marrlll, Schneider, & Fletcher, 1980) has attended to other kinds of variables and adopted a “systems” orientation (Dick, 1987) involving testing in more authentic contexts and the use of multiple dependent variables. Throughout its history, the tradition has favored technology- driven research in whlchthe emergence of some form of technology (e. g., microcomputers [More & Ralph, KHZ], hypertext, CD ROMs [Riding & Chambers, 1992)) gttmulates a research to evaluate its effects on learning outcomes.

Though CA1 is the oldest paradigm for work in IT, it is by no means an abandoned one. Applications designed under this paradigm range from early drill- and- practice programs to more recent network- based World Wide Web documents. 5 They account for the bulk of instructional software now in actual classroom use, and evaluating the instructional effects of such appli- cations continues to be an active area of research.

.!? S Wyhtm The emergence of the next paradigm was the direct result of an immigration, which began in the early 1970% of workers from the field of Artificial lntel-

‘1 by no means wish to suggest by this that o/ l Web applications should be viewed as extensions of the CAI paradigm. The World Wide Web is very much a work in progress and I Only wish to observe that at least some of its current applications, in their design and methodologies of evaluation. are consistent with the traditions of CAI research.

i 1. PARADIGM SHIFTS AND INSTRUCTIONAL TECHNOLOGY 7 ,ffgence (AD researchlhto the educational arena. Carbonell’s thesis defense ‘( 1970) was cited by Wenger (1987’) as the event that marked the onset of this influx. Research in AlIs founded upon the conjecture that cognition 1st ij~ some sense, a corn ‘Qn& ictlon of %rteni onal process that can be studied through the”,

systems that serve as functional models of the otherwise inaccessible processes of the human mind (Pylyahyn, 1989)... ffP

machines can be programmed to display intelligent behavior, there is,. noz reason, at least in principle, that systems could not be designed to assume the role of a stilled teacher. Since oneonone tutoring is commonly consid- ered the gold standard against which other methods of instruction are measured (Bloom, 1984) the paradigm is founded on the proposition that education could be globally improved by providing every student with a personal (albeit machine- baaed) tutor (Lepper, Woolverton, Mumme, &Curt- ner, 1993).

Information F’rocessing Theory (Simon, 1979) served as one of the found- ing premises for work in AI. It held that problem solving (human and oth- erwise) could be seen as a process of defining a representation of a problem space consisting of an initial state, a goal state, and a set of operations for moving from one state to another. By this view, representation became a central issue for understanding both problem solving and cognition in gen- eral. Learning, in this light, becomes the process by which the problem solver acquires a proper representation of a problem space. instruction, then, consists of activities designed to facilitate the acquisition of such a representation by the learner. The role of technology in this process is really not so different from the role that it assumes within the CAl paradigm. The differences are more in degree than in kind. In both cases, the’designed application serves instruction by posing problems and by providing feed- back to the learner. The difference is that ITSs aspire to do this in a more interactfve fashion and with respect to a more complex set of skills.

Much more striking differences are seen, however, in the evahratfve methods which comprise the paradigms. Unlike the CAf paradigm which reflects the standards and methods of the general educational research community. the 1TS paradigm applies an approach adopted from research in AI. Al research is dedicated to the task of providing an accoun@ ln computationaf terms (f. e, algorithms and representational schemes), of various aspects of human cognition. The process by which this is accom- plished was described by mnert (1984) as follows:

1. Propose a theory &explain the phenomenon. 2. fmpfement the theory In a computer program designed to simulate the

phenomenon. 3. Run the program. 4. Analyze the program’s output. (p. 24)

8 KOSCHMANN When l refer to the ITS paradigm, therefore, I am referring to work that applies the methods of Al research to the task of understanding skilled tutoring in complex domains. Competent tutoring in such domains raises several problems in representation- how to represent the knowledge of an expert in the domain, how to represent the pedagogical expertise of the tutor. and how to represent the (possibly faulty) understanding of the student user (Wenger, 1987).

Research conducted under this paradigm leads to the generation of a different,@ of research questions from those addressed within the CAl tradition. Whereas Instructional efficacy is the sine qua nofl for CAI re- searchers, the critical issue for ITS researchers is inslmctiurtul competence; that is, does the application faithfully emulate the behavior of a skilled tutor? The focus, therefore, is on the fidelity of the system’s performance, rather than tts effect on student learning outcomes. 6 This shift in priorities has been a source of misunderstanding among researchers working within the two paradigms. To an ITS researcher, a completed program serves as an existence oroof for a theory, whereas to a CAl researcher, no project is complete until ;he application’s value has been demonstrated in the classroom.

b In the end, however, these two paradigms have more in common than is usually appreciated. Although one is implicitly behavioristic in its approach

and the other explicitly cognitive, both assume an epistemologicaf stance that is realist and absolutist (Doerry. 1994, Ernest, 1995). Both reflect pre- “vailing notions of knowledge as given and of teachers as the final authority (Schommer, 1999). There is an implicit commitment to the exfstence of a “correct” representation and a view of the tutor as an agent for effecting “the learner’s acquisition of this representation. Furthermore, like the CAf

developers before them, ITS researchers embrace a rather conventional Tvtew of teachfng as delivery, what has been termed a bansmission model of instruction (Pea, chapter 7, this volume). Wenger (1987), for example, argued that “the ability to cause and/ or support the acquisition of one’s knowledge by someone else, via a restricted set of communication operations” was the cantral problem of ITS design (p. 7). As we see, however, later paradigms represent a departure from these received norms, both in their underlying epistemological frame of reference and in their models of instruction.

-1s is not to say that there has been no research on the efficacy of Intelligent Tutoring Systems. However, mwt research within the ITS paradigm (as I have defined it here) has concerned itself with issues other than efficacy (e. g., what accounts for expertise (Koedinger &Anderson, 19901, how to provide plausible explanations to the student [Clancey, 19831. how to represent the student’s faculty understanding (VanLehn. 19821, the pragmatics of student/ tutor Interaction [Woolf & McDonald, 1984)). Although recent research in instructional design (e. g., ‘structural learning’ (Scandura & Scandura. 1988], ID2 [Merrill, Lin. & Jones. 19901) is reminiscent of earlier lTS work in Its emphasis 011 knowledge representation, its behavioristic evaluative traditions align it more comfortably with the CAI paradigm.

1. PARADIGM SHIFTS AND INSTRUCTIONAL TECHNOLOGY Logo- as- Latin Paradigm

9 The next paradigm arose from an epistemologicaf perspective that holds knowledge to be acquired through “a process of subjective construction on the part of the experiencing organism rather than a discovering of ontologf- cal reality” (von Glasersfeld, 1979, p. 109). This view of learning, which is explicitly relativistic and fallibilist (Ernest, 1995). is referred to as construe- tiukm. ’ It had its origins in the work of the developmental psychologist Piaget, who introduced a theory of learning whereby new information inter- acts with prior knowledge through a process of assimilation and accommo- dation (Pfaget, 1985). This constructivist view of learning inspired the development of a number of instructional methods (e. g., “learning by dis- covery” [Shulman & Keisler, 1966]; Open- Classroom Learning, [Kohl, 19691; Experiential Learning, [Kolb, 19841: Inquiry Learning [Bateman, 1999]) all dedicated to the proposition that learning occurs most propitiously under circumstances of personal inquiry and discovery.

Papert (1980) argued that the activity of programming computers could play an important role in constructivist learning.* Computer programs are particularly interesting artifacts for a learner to construct because, unlike term papers and other traditional class projects, they are executable. In building an executable artifact, such as a microworld or a computer- based simulation, the learner in effect “teaches” the computer, thus providing a new role for technology in learning. Instead of serving as a stand- in for the teacher, as was the case in the CAf and ITS paradigms, the computer becomes “tutee” (Taylor, 1980) allowing the learner to assume the role of teacher. The assumption here is that by engaging in the activities of programming- design- ing. building, and debugging programs-- the learner acquires cognitive bene- fits that extend beyond simply learning to code in a particular language. A substantial research literature has accumulated that addresses the question of just what these benefits might be (Mayer, 1988; Palumbo, 1990; Pea & Kurland, 1987; Salomon & Perkins, 1987). Much of this research involves learning to program in Logo, a powerful programming language designed by Wafly Feurzeig in the mid- 1960s for use by young children (Papert, 1989). Because much of this work focuses on learning to program jn the service of more general educational objectives, I have termed this research approach the Lcgous- Latin Pumdkm (Koschmann, in press).

‘This is admittedly B blt of a glosswonstructivism is more a shared orientation than a unified school of thought. Within the community of workers collectively labeled as “coostruc- tfvists” can be found a number of competing perspectives including mdkal conslruclivism (van Cfa?. ersfeld, 1979). ecokgicolconstrucfiuism (St& r. 1995). sniolconslrucfitim (Bauersfeld, 1995), and advocates of Cognitive Flexlbfllty Theory (see chapter 2. this volume), sometimes labeled hkmofionprwessing consfrucriuisb (Steffe & Gale, 1995).

‘Because Of its important role in stimulating later research. I have selected the publication Of Papert’s Mindstorms as the inaugural event for the emergence of this paradigm.

KOSCHMANN IO

Exploring the cognitive benefits of programming can be seen as one part of a broader movement in educational psychology to identify mechanisms lor fostering the development of general skills for learning and problem- solving (Bruer, 1993; Segal, Chipman & Glaser. 1995). As a consequence, researchers working wlthln this paradigm have utilized the standard re search methods of educational psychology in assessing the cognitive bene fits of learning to program. Whereas research under the CAI Paradigm is concerned with instructional efficacy, Logo- as- Latin research focuses more specifically on the Issue of instmcticfd tmns~ er. programming instruction is treated as the experimental intervention, and subsequent performance on other related tasks serves as the dependent variable. The use of control groups is common. Studies, so constructed, have investigated the effect of learning to program on planning (De Corte, Verschaffel, & Schrooten, 1992). metacognition (Clements & CUIIO, 1984) and other aspects of cognitive performance (Lehrer & Littlefield, 1993):

Constructivist research takes as its central concern the issue of cognitive s& organbatlon (Cobb, 1994). In so doing, it adopts the view of mind as a ‘ phenomenon residing within the head of the indlvfdual. This is a view that ls deeply steeped in western philosophical traditions and that Is founda- tlonal to most current research in psychology and education. It is not universally held, however. ‘f? rere, are competing views that place the mind wlthfn the surrounding sociocultural environment. As we see in the next section, these alternate views have important implications for educatron and the use of technology therein.

CSCL: AN EMERGING PARADIGM IN INSTRUCTIONAL TECHNOLOGY

I argue in this section that we are currently witnessing the emergence of a snew paradigm in IT research; one that is based on different assumptions about the nature of learning and one that incorporates a new set of research practices. Although there is a noted lack of agreement among the previously described paradigms with respect to their theories of learning and pedagogy, all three approach learnlng and instruction as psychological matters (be they viewed behavioristically or cognitively) and, as such, are researchable

Yft IS worth noting that not all Logoas- Latin research is based on Logo; nor does all research fnvofvfog programmfng in Logo necessarily represent Log- as- Latin research. There have been, for example, related studies exploring the cognitive benefits of programming in Prolog (Schers, Goldberg, &Fund. 1990, Verxmi & Swan. 1995). Conversely. there is considerable research using Logo that is not concerned with the issue of transfer. This is true. for example. of much of the research done by Papert and hls associates( e. g., Harel& Papert, 1991). Following in the tradition of classfcal Pfagetfao research, much of Papert’s work with Logo has tended to consist of case studies designed to document children’s achievements while working with Computers.

1. PARADIGM SHIFTS AND INSTRUCTiONAL TECHNOLOGY II by the tradftional methods of psychological experlmentatlorf. This newly emerging paradigm, on the other hand, is built upon the research traditions of those dlsclpffnes- an$ hropology, sociology, lingufstics, commumcatlon sdence- that are devoted to understanding language, culture, and other aspects of the social setting (cf. Scott, Cole, & Engel, 1992). As a result; tBs reflects a different view Of learning and instruction, one that brings these social issues into the foreground as the central phenomena for study (Hutchins. 1993). This perspective has been influenced by a number of recent movements in the socially oriented (as opposed to the psychological) scl- ences. I briefly describe three, although there were certainly others that have contributed to this Zeitgeist. lo

Socially Oriented Constructivist Viewpoints cO~ S~* Ucthkn orlginahy arose out of f’iaget’s research in developmental psychology and has developed into an important perspective in educational research (cf. Steffe & Gale, 1995). Within the constructivist camp, there Is a growing Interest In the social context within which learnfng occurs. Notable in this regard Is the research of the so- called wo@ ons, who have emphasized the importance of peer interaction for cognitive development @Oise & Mugny, 1984). In educational research (particularly in mathematics education), a school of thought known as social constructiuis@ as emerged (Bauersfeld, 1995; Cobb,, 1994). As a constructivlst perspective, it takes a nonabsolutist, fafllbllist view of knowledge as constructed, but, unlike other

constructivfst positions, views this construction to be an essentially sor- jd process (Ernest, 1995).

Soviet Sociocultural Theories Another important Influence was the research of Soviet psychologists inter- ested In the cultural basis of human intellect. Perhaps the best known of these was Vygotsky, who formulated the theory of cultural- historkal psychd- ogv (van der Veer & Valsiner, 1991). His Geneml Genetic Law OF Cuftural

“‘ Two other movements not discussed here but worthy of mention are Symbolic fnterac. tfonfsm and Social Constructionkm. Symbolic lnteractionism has its roots fn the writings of the American Pragmatfst philosophers. particularly George Herbert Mead @turner, 1969). .& ao anafytfc framework. however. it shares many of the concerns of the other approaches described here, esp% faffY the Soviet socfocoftural theories and Situated Cognition (Star, 1%). %ciaf Consttwtfoofsm is another related movement that represents a research tradftfon in socfaf PsYC~ fogY and SoCiofogY (Gergen, 1985; Ha& 1986). Constructkxdsm (the “N” word rather than the ‘V” word) k dedicated *to the task of describing what the ‘inner fife of a ‘ffngufstfcaffy situated Person’ in a socially constructed world fs like” (Shotter, 1993, p. 161). Evidence of thfs ‘“ oer fife fs extracted from the study of day- today communicative activities, discursive practices. rhetoric. and argumentation (Billig, 1987). Social Constructionists. like the sociaffy oriented Cowtructffists. are explicitly nonabsolutist in their views of the nature of knowledge.

A

I2 KOSCHMANN Dex=+ pmenf stipulates that learning always occurs on two planes: first on the inter- psychological, and only later on the intra- psychological (Wertsch, 1985). As a mechanism for learning on the inter- psychological plane, Vygot- .& J hypothesized the existence of a construct that he termed the z~ 0. e d, p& ma/ &twdqmt? nt (Vygotsky, 1978). This zone represents the enhanced capabilities of a learner working in the presence of a more skilled coworker or teacher.

The cultural- historical approach to learning developed by Vygotsky fo cused largely on the role of language in intellectual development (Brushlin- sky, 1990). A related school, represented most prominently by the Russian researchers Leont’ev (1974) Galperin (1992). and Rubenstein (Brushlinsky, 1989) focused its attention on the role of activity in human development.” One articulation of the so- called Activity Theory, attributed to Rubenstein (Brushlinsky, 1999). asserts that “The subject not only reveals and manifests himself in his actions and in the acts of his independent creative activity: he is created and defined in them. That is why the things he does can be used to determine and mould his character” (p. 67). Activity Theory takes, as its unit of analysis, human goaldirected activity in Its cultural context (Leont’ey, 1974). It focuses, therefore, on signs, symbols, rules, methods, instruments, and other artifacts that serve to mediate this activity.

Vygotsky’s cultural- historical psychology and the work of the later Activ- ity Theorists has subsequently developed a following both in educational research (Forman & Caxden, 198% Griffin & Cole, 1987; Newman, Griffin, & Cole, 1989) and in the specialized area of computer science dealing with human/ computer interaction (Kuuti, 1996).

Theories of Situated Cognition The term s& rated, as in “situated learning” or “situated cognition,” has assumed a variety of meanings in different disciplinary contexts. It refers to a specific theory in linguistics and philosophy of language (Barwise &Perry, 1983) a reaction in the Al community to symbolic models of cognition (Clancey, 1993; Winograd &Flores, 1986) a program of study in anthropology (Suchman, 1987) and a way 01 reconceptualizing educational practice (Brown, Collins, & Duguld, 1989; Creeno, 1989; Lave & Wenger, 1991). It is the latter two senses that concern us most directly here. In theories of situated cognition, learning is viewed as a process of entry into a community

“The Russian dyeroryefnasr is commonly translated into English as “activity.” Many Russian scholars, however, are not completely comfortable with this translation. German has two words. Ahtiuitiland TiMgkeit, that both translate to ‘activity.” The latter is composed from the adjective fiitig. meaning busy or engaged. It is used in expressions such as in Ttitigkeit seken. meaning to engage or put into action. Consequently. this term comes closer to capturing the meaning

I of the Russlan dyqalyekmsr than the usual English translation. 1, PARADIGM SHIFTS AND INSTRUCTIONAL TECHNOLOGY 13

of’pratiicft, to wit: “To learn to use tools as practitioners use them, a student, like an apprentice, must enter that community and its culture. Thus in a significant way, learning is, we believe, a process of enculturation” (Brown, Co]] jns, & Duguid, 1989, p. 33). Within this perspective. the context (both social and material) within which learning occurs comes under careful scru- tiny, arising from a view “that agent, activity, and the world mutually con- stitute each other” (Lave & Wenger, 1991, p. 33).

Taken together these perspectives- social constructivism, Soviet so- ciocultural theories, and situated cognition- provide the intellectual heritage from which CSCL has emerged as a new paradigm for research in instruc- tional technology. Although they arise within different disciplines and utilize different metaphors of social process (Gee& z, 1980) they all represent a gestalt- like shift in point of reference relative to the views taken by the paradigms described previously. This shift in point of reference, leading to a foregrounding of the social and cultural context as the object of study, produces an incommensurability in theory and practice relative to the para- digms that have come before. _, .

The model of instruction underlying work in CSCL is termed “collaborative learning.” Although it is easy to recognize examples of collaborative learning, it is difficult to provide a precise definition. Bruffee (1993) described it as -a reculturative process that helps students become members of knowledge communitieswhosecommon propertyis different from thecommon property of the knowledge communities they already belong to” (p. 3). This definition, focusing on what collaborative learning is meant to accomplish, resonates with the view of learning as entry into a community of practice. On the other hand, Roschelle and Behrend (1995) described it as “the mutual engagement of participants in a coordinated effort to solve [a] problem together” (p. 70). This latter definition highlights several facets of the method: a commitment to learning through doing, the engagement of learners in the cooperative (as opposed to competitive) pursuit of knowledge, the transitioningof the instruc- tor’s role from authority and chief source of information to facilitator and resource guide. Examples of collaborative learning methods include Expedi- tionary Learning,‘ 2 Group Investigation (Sharan, 1980) Problem- Based Learn- ing (Barrows, 1994; Barrows & Tamblyn, 1980; Koschmann, Kelson, Feltovich, &Barrows, chapter 4, this volume), Project- Based Learning (Blumenfeld et al., 1991; Soloway, Krajcik, Blumenfeld, & Marx, chapter 11. this volume), and other forms of small- group learning (Noddings, 1989; Webb, 1982).

Over time, interest has grown in the question of how technology might serve to support collaborative methods of fnstruction (Crook, 1994; Kosch- mann, 1994a). There have been a number of significant events germane to the emergence of this area of work as a new paradigm in IT. A preliminary

“A method utilized in a New American Schools Development Corporation (NASDC) project undertaken by Outward Bound.

KOSCHMANN 14

exploration of the issues engendered by the useof technology in collaborative education took place in 1983 at the Conference on Joint Problem Solving and Microcomputers held at the Laboratory of Comparative Human Cognition (~ HC; Cole, Miyake, & Newman. 1983). A later workshop, conducted under the auspices of the NA’IYO Special Program on Advanced Educational Technol- ogy wa. s held in Acquafredda di Maratea. Italy in 1989 (O’Malley. 1995). Because this was the ffrst gathering to adopt the title “computer- supported collaboratfvel~~” I have chosen this event to mark the emergence of the paradigm. Subsequent CSCL workshops were held, one in 1991 at Southern fllfnofs University (Koschmann, 1992) and another at Ontario lnstitute for Studies fn Education (OISE) in 1992 (Koschmann, Newman, Woodruff, Pea, & Rowley, 1993). The first international conference on this toptc took Place at the University of Indiana in the fall of 1995 (Schnase & Cunnius. 1995) and a follow- up is planned for 1997 at the University of Toronto.

As reflected in the chapters of this volume, CSCL applications assume a variety of forms. They can be categorized on a number of dtmensrons, including the focus of use, how the use is coordinated in time, and the instructional rola. it was designed to serve. Though the majority of CSCL applications are designed for student use, there is also a need for tools to support teachers engaged in collaborative forms of instructron (see chapter 11, chapter 5). The locus of use may be intra-, inter-, or e~- ChSSrOOm (Koschmann & O’Mafley, 1994). Applications have been desrgned for use wfthln the classroom (chapter 9. chapter 4. this volume), to connect users across classrooms (chapter 8), and in some cases to create “virtual CRASS- ‘: rooms: (Hiltz, 19%). Users of an application may coordinate their interaction

sym& onously (e. g., chat programs) or asynchronously (e. g., e- mail). CSCL applications may serve a number of roles. Technology, for example, can be used to present or simulate a problem for study, helping to situate it in a real- world context (e. g., chapter 4, this volume). Alternatively, computers . . can be used to mediate communication withm (chapter 6) and. across classrooms (chapter 8, chapter 5) or to introduce new resources mto the classroom (chapter 7). Computers can also provide archival storage for the products of group work, thereby supporting “knowledge building” (chapter 10). Finally, computers can support the creation of representational formal- isms that enable learners to model their shared understanding of new concepts (e. g., the Envisioning Machine described in chapter 9).

Unlike the types of issues (i. e., instructional efficacy, instructional compe- tence, instructional transfer) underlying the paradigms described earlier, research in CSCL is concerned with questions such as how IS learmng reflected in the language of learners (chapter 9)? How do social factors enter into the process of fearhing (chapter 3)? How is technology actually used in collaborative settings (chapter S)? Stated differently, the central focus for ‘research in CSCL is on instructitm as enacted pmctice. Consistent with the

,. PARADIGM SHIFTS AND INSTRUCTIONAL TECHNOLOGY I5 sociocufturaf outlook of its practitioners, research in CSCL tends to utilize the research methods of the sO& tf sciences (for more on this see chapter 7, this volume). Although the paradigm is still very much in its formative stages, several comments can be made concerning the general analytic framework of research in this area. First,, driven by the types of research questions being asked, work in CSCL te$ s~ to. focus on process rather than outcome. Second, there is a central concern with grounding theories in observationa. f data (Chaser &Strauss, 1967) and in the construction of thick descriptions (Cuba& Lincoln, 198f) ofthephenomenaunderstudy. Asaconsequence, CSCLstudfes tend to be descriptive rather than experimental. A third and final aspect of this emerging body of research is that there is an expressed interest fn understanding the Process from a participant’s viewpoint. As argued by Jordan and Henderson (D95), learning can beat be understood “as a dfstrfb uted. OngOing Social Prom% where evfdence that learning is occurring or has occurred must be found In understanding the ways in which people coffab natively do learning and do recqgnizing learning as having occurred” (p. 42, italics added). CSCL research focuses, therefore, on participants’ talk, the artifacts that suPPort and are produced by a team of learners, and the partfcf- pants’ own accounts of their work. There are a smafLput @owing number of studies that fit this descript& (cf. chapter 9, this volume; ~Gi& n, Koschmann, &Conlee, 1995: Griffin, Befyaeva, & Soldatova, 1992; Roth, in press).

It should be acknowledged that while all of the chapters in this book describe work at the confluence of technology and classroom collaboration, not alf neCesSarilY espouse a social theory of learning, nor do they all speak to the research issue of instruction as enacted practice. Although this may appear problematic given the description of the paradigm provided here, f think there are a number of ways of accounting for this discrepancy. One possibility, for example, is that some of the current researchers in the area continue to be influenced in their work by past paradigms; that is, that they CUrrentIY exist with a foot in both worlds. This seems quite plausible, given the relative newness of the paradigm. Another possibility is that there may be more than one paradfgm emerging with a commitment to collaborative forms of instruction. fn addition to the paradigm described here, there may be one or more other paradigms with a more cognitive orientation. ft is difficult to know for sure. In the end, it is always easier to provide an account of paradigms Past than it is to describe a paradigm in the process of becoming.

LOOKING INTO THE FUTURE: HEGEL VERSUS KUHN

The lour paradigms described in the chapter are summarized in Table 1.1. No Claim fs made that this list is necessarily exhaustive. Indeed, it is recog- nized that there are examples of IT research that do not fit within any of

16 ,. PARADIGM SHIFI’S AND INSTRUCTIONAL TECHNOLOGY 17

the paradigms described. Some of this work may be anomalous and does not subscribe to any particular paradigm, but the point is readily conceded that there probably exist additional paradigms that have not been discussed hereJ3

The analysis offered in this chapter provides a new scheme for catego- rizing work in IT. There have been numerous past attempts to create tax- anomies based on the role that the application was designed to play in the instructional setting (Soloway, 1993; Taylor, 1980; Wu, 1993). Taylor’s (1980) typology of tutor, Wee, and t* is probably the best known and is one that has been adopted by a number of other authors (Crook, 1994; Dreyfus & Dreyfus, 1986; O’Shea & Self, 1983). It appears to have several weaknesses, however. By focusing exclusively on the functional nature of the application, opportunities to consider other aspects of the work- such as the theories of learning that motivated it in the first place- are missed. Second, by trying to reduce the diverse set of IT applications into just three categories, con- siderable resolution is lost. Although more elaborate typologies have been proposed (e. g., Wu, 1993). it is not clear that this is the best direction to be taken. By focusing exclusively on descriptive aspects of the application, we lose the ability to discern larger shifts in philosophy and practice. By con- trast, applying a ffihnlan ‘analpls encourages a broader, vfew of practice, one that encompasses underlying theories and methods of research’zd argumentation.

Various authors have made attempts to divine the direction that IT research might take in the future. In many cases, this is done in the form of a dialectical analysis. This method, developed by the Nineteenth Century philosopher Hegel, is based on the theory that our understanding of a concept proceeds through a@ h&+ part process of clarification- a the& s is opposed by its antithesis and is eventually supplanted by a new synthesis (Koschmann, 1994b). For example. Larkin and Chabay (1992) and Duffy and .Jonassen (1992) contrasted work in the CAI and ITS traditions in the interest of identifying possible directions for future work. Derry and LaJoie (1993) focused on the contrast between ITS and constructivist- motivated research and argued that future work would represent a synthesis of these two approaches. Most recently, Cobb (1994) Crook (1994). and Steffe and Gale

‘brie candidate that comes immediately to mind is research related to ‘CSCWritktg” (Gruber. Bruce, & Peyton. 1995). There is a substantial body of work devoted to the use of CMDpttterS In composltfon (see the Neuwirth and Wojahn chapter for references) that is largely invisible to the IT community because it is embedded in the llteratwe of writing Insttuctlon. The question of whether CSCWrltlng’should be vlewed as a special dfsclplkwy interest withln CKL or as a paradigm in Its own right does not have a clear answer at this point. What is Clear, however, is that the two movements share many issues and that there is much that researchers in CSCL could learn from the accumulated experience of the composition community.

KOSCHMANN IS

(1995) have contrasted constructivist and sociocultural views Of learning in the hopes of achieving some form of reconciliation.

The historiographic account presented in this chapter makes this dialec- tic- approach problematic, however. In no case did a newly emerging, paradigm PP a ear to be the synthesis of ideas drawn from previous para- digms. The lTS paradigm was less an adaptation of prior work in CAI re-

search than an fnv& on of a new group of workers bringing with them new standards for design and evaluation. Similarly, the Logoas- Latin paradigm was not presaged by the CAl or ITS paradigm; it represented an entirely different Philosophy about the use of technology in education. Finally, the emergence of the CSCL paradigm could have been in no way predicted by the clash of constructivist and information processing theories of learning.

Ironically, the ultimate lesson of this form of analysis is that the revolu- tionary changes that Kuhn described as paradigm shifts are always difficult &to’ for- and, in particular, cannot be adduced from the study of past

history. The ideas that have shaped work in IT, have, in general, come from ou& fde. ae field. As a result, the task of identifying the sources of future shifts is a difficult one. Kuhn, himself, despaired at the prospect of ever providing a complete account of how a field- defining, revolutionary idea comes to exist. He lamented, ‘What the nature of that final stage is- how. a+ slndlvldual invents (or finds he has invented) a new way of gfving order tq idata now all assembled- must here remafn inscrutable and may be perma-

nently so” (1972, p. 90). And so it may be for our own efforts to foretell the future direction of research in instructional technology.

ACKNOWLEDGMENTS The author would like to thank Paul Feltovich and Alan Lesgold for reading an earlier draft of this chapter and providing many constructive comments. The author was supported by the Spencer Post- Doctoral Fellowship from the National Academy of Education while preparinq this chapter.

REFERENCES errows, H. S. (19%). Pmcticpbased learning Problem4xrosed learning applied 10 medical edwodo”.

Springfield, IL: Southern lllinois University School of Medicine. H. s., & -~ II~ IY”, ft. (IWJ). R& em& sed learning: An appmrch ro medical education. New York: Springer. Bar- wise, J., &perry, J. (1983). Sihmrions and offit” des. Cambridge. MA: MIT Press. Bateman, w. I.. #wO). open to ,pehm: The or! of teaching and learning by inquiry. Se” Francisco.

1. PARADIGM SHIFTS AND INSTRUCTlONAL TECHNOLOGY 19 Bauerefefd. H.( l99.5). The StruCturing Of structures: Development and function of mathematizlng

as e So& l practice. I” L. P. Steffe &J. Gale (Eds.). Caw/ ructitim in education (pp. 137- 158). Hillsdale. NJ: Lawrence Erlbaum Associates. Biagloli. M. (199( I). The anthropology of incommensurability. Studies in Histov and Ph;/ osophy

dsdem?, 21. 183- 209. Billig. M. (1987). Arguing and fhinbing A rheforicol opprmch to socio/ ps~ hd~. Cambridge, “I(:

Cambridge University Press. Blafedefl, F. J. (1976). Historical development of computer- assisted instruction. ~wrnof of Ed&

cotionol Techndqv Systems, 5, 155- 170. Bloom B. S. (1984). The 2 Sigma problem: The search for methods of group instruction ee

ekctk es one- km” e tutoring. Educrzio” ol Researcher, 13( 6), 4- 16. Blume” feJd. P.. Sofowe~. E. Marx. R.. Kraicik, J, Cuzdial. M.. & Palincsar. A. (1991). Motivating

proiect- based learning: Sustaining the doing, supporting the learning. Educotioi psFbd,, gisf, 26. 36* 398. Blunw H. (1969). Symbdic i” feracfionirm. Englewood cliffs. NJ: Prentice- Hell. Brow”. J. S.. Collfns, A.. & Duguid. P. (1989). Situated cognition and the culture of IearnIng.

Educational Researcher, 1x1). 32- 42. Bruer. J. T. (1993). Schools for thhwghl. Cambridge, MA: MIT Press. Bruffee. K. (1993). ~o( loborolioe learning. Baltimore: Johns Hopkins U” iverstty press. Brushfinsky, A. V. (1989). Sergei Rublnstein: Founder of the activity approach in psychology.

Soviet Journal ofPsychdogy IO. 24- 42. Bruehff” sky. A. V. (1990). The activity of the subject and psychic activity. h, v. A. ~ektovsky, &

Y. hgeetr6m We.). Acfioity. %wies, metbed&@, andpmb/ ems (pp. 67- 122). Orlando. n: Paul M. Deutsch Press. CarbonelI. J. (1970). Mixed- initiative man- computer instructional dialogues (Doctoral d& sert,+

tie”, .bsxhusetts 1” stlttrte of Technology, 1970). American Doclom/ Dissertodo,, s, 10. 62. Cfance~. W. (1983). The epistemology of a rule- based expert system: A framework for expfene- $

tion. Artificial Intelligence. 20. 215- 252. ; i

CfenceY. W. (1993). Situated action: A neuropsychological interpretation. Cq” itiw Science, 17. 87- 116. Clement% D. H.. & Gullo. D. F. (1984). Effects of computer programming on your children’s

Cognition. JOumd ofEducational Aychdqgy. 76, 1051- 1058. Cobb. P. (1994). Where 1s the mind? Constructhat and socioculturai perspectives on methe-

matical development. Educatio” al Researcher. 2317). 13- 20. .

Cole M., W= k=. 1’4.. & Newman. D. (Eds.). (1983). Pmceedi” gs offhe Conference o” point problem Sduing and hkWOmP” ferS (Tech. Rep. NO. 1). La Jolla: University of California. Se” Diego, Laboratory of Comparative Human Cognition. (ERIC # W23839n Coufse”. J.. ‘& teva”, D. P.. Melaragno, R.. & Sllberman, H. F. (1962). Effects of branching in r, iIcomp” ter~ Co” trolfed auto- instructional device. burnol, ofA C ok. C. (19

If f lied Psychology, 46,,~~~~~

Kegan Pa I. ). Cwnp~ ler~ and the cdlabomtiue experien e o earning. London: b C a”. L. (19 ). Teachers and machines: The classroom us

Teachers College Press. rf I

oft hndw since 1920. Cuban. I.. (1993). How leachers taught. New York: Teachers College Press.

” CYU’~ C I 1SJJ. De Carte, E., Verschaffer L., & Schroote”. H. (1992). Cognitive effects of learning to program in itivr Pffwr~ nf learning to program in Logo: A one- Year study with sixth graders. In E. De Carte, M. Linn. H. Mandel. & L. Verschaffel rla” de. 1. & L. Verschaffel @de.). COmPWer- boSed leorni+? envimnmenh and problem soloing (pp. 20% 228). Heidelberg:

Springer- Verlag. 4.. z5 ,,+. 20% 228). Heidelberg: Derr~, S. J.. & Lejoie. S. P. (1993). A middle camp for (un) intelligent instructional computing. I”

S. P. Leioie. & S. J. Derry( Eds.). Compufers os cognitive fools (pp. I- 11). Hillsdale. NJ: Lawrence Erlbaum Associates.

KOSCHMANN 20

Dick w, (, gg7), A higq of instructional design and its impact on educational psychology. fo J: A, ~~~~~~~ & R. R. Ronning (Ed%), Hisloffcaf foundations of educolionol PsYCholosV (PP. p& 200). NW York: Plenum. Doerry E, (fggq. &@ ng for lxdrwk: A casual inspection d Ihe episremdogfcof foundations of

,hree dfs& f;,, es (Tech Rep. NO. CiS- TR- 94- 25). Eugene: University Of Oregon. Department of compute, and Information SfieoCe. Doise W,, & f,, tugny, G. (1984). the social deoefopmenr of rhe inteUec(. Oxford: Pergamoo. Dreyfhs, H,, & Dreyfus, S. (1966). Mind owr machine. New York: Macmiiian. Dulby. T,, & ~~~~~~~~ D. (1992). Consnuctiursm and the IechndqgY d inShUCdOn. A ConeersQhm.

Hill& les NJ: Lzw~~ IW~ EIlbaom ASSOCiateS. Er,, est, p.(, m), The one and the many. In L. P. Sreffe & J. Gale (Eds.), COnslrucIf~~ Sm fn edocoffon

Forma,,, E,, & Carden, C. (1985). Exploring Vygotskian perspectives in education: The cognitive cpp, 459- 486). Hillsdale. NJ: Lawrence Eribaum Associates.

vafue of peer interaction. lo J. Wertsch (Ed.), Culture, communicadon andcognition Vygorskian K, ir,! es (pp. 323- 347). Cambridge. UK: Cambridge University Press. ~$ f:, M. (~~ 8). karning hierarchies. Educafiond Pvhologisl 6. I- 9. ~i~ e; k, p. (1992). The problem of activity in soviet psychology. J~ umaf ofR~ s. sfon and f% r

~,,,& on Psychology, 30,37- 59. ~eertz. C. (1980). Blurred genres: The reffguration of social thought. American Schdar. 49. 165- _~

179. c. rrr, en. J. (198~). The so& i constructionist movement in modern psychology. American Ps~

--~~~ o; c@ s;, 40. 266- 275. ~~ i,,,~ o D, (1~ 7). Feedback, prornptlog. and overt correction procedures in non- branching

co; put~ r. a&, ed ,m. t, uctlon programs. Journal of Educadonol Research 60. a- 426. clever, B, &Strauss, A. (~$ 67). The discowry of grounded theory. Chicago: Aldine- Atherton, GIN”,, p. Koschmann, I’, & CO~ Z, M. (1995). Theory sequences in a problem- based learning

g,& A case study. In pmceedings of the Firsl fn~ errmfion0f ConfePZCe M Cmpeler SuPPod for Coffobomf+ e Leoming @CL 951 (pp. 139- 142). Mahwah, NJ: Lawrence Erlbaom A~ soct- ates. ~~~~~~ J. c, (KXXI). Shuations, mental models and generative knowledge. lo D. Klahr & K.

&., ksy( Eds.), Cwnplexinfwmorionprocessing: The impoCl ofHerberlA. Simon (PP. 285- 318). Hillsdale, NJ: Lawrence Erlbaum A?. SoCiateS. Grirfio P Relyaeva, A, &Sold& ova, G. (1992). Socio- historical concepts applied to obsewatfoos

of ioLputer we. Ee~ opea~ .f~ umol of Psychdosy of Educadw 7, 269- 286. ~, gfi,, p. & cole M. (1987). New technologies, basic skills. and the underside of education:

&& to be done? ln J. A. Laoger (Ed.), Lunguoge, litererocy, ond cufmre: fssues of SmierY and r& ofing (pp. 199- 231). Norwood. NJ: Abiex PubliShi% ~r” kr S. Sruce, B. C., & peytoo, J. K. (1995). Collaborative writing in multiple discourse Coo-

text;. &m, ww Suppwred Cmpemh’w Work, 3, 247- 269. rJuba, E., & Lkwxb~, Y. (1981). Effecriw euoluorion. San Francisco. CA: Josse~- Rass. &rel ]. & papert S. (Ed%). (1991). Conshwfionism. Norwood. NJ: Ablex. Harr; d. (198~ A; outline of the social constructionist viewpoint. In R. Ha& (Ed.). The SC&?/

&&,,, I of enmtions (pp. 2- 14). Oxford, UK: Black~ ell. HRtz S. R (w& 3). Collaborative learning in a virtual classroom: Highlights of findings. lo m

ckedings of 1988 ACM Conference on Computer- SuppOred Cooperodw Work (PP. 282- 290). New York: ACM Press. Hutchins, E. (vX. B). Review of Sihroted learning: Legilimole peripheral participation. American

Andrropdc@ sl, 95, 743- 744. ~0”~ s M. &vane P. (Eds.). (1992). Adoptioe learning enuironmen~ s. New York: Springer- Verlag, Jo, d& i. & Henderson, A. (1995). Interaction analysis: Foundations and practice. Jownaf @‘

Ihe’ledming Sciences, 4.3% 103. KLtcher, p. (1978). Theories, theorists. and theoretical change. PhiroSophicOr Review, 87. 519- 547.

1. PARADIGM SHlf% AND INSTRUCTIONAL TECHNOLOGY 21

Koedhger. K.. & Anderson. J. (1990). Abstract planning and perceptual chunks: Efemeots of expertise in gmwtry. CognitiueScience, 14, 511- 550. Kohl. H. R. (1969). The open cfossror~ n. New York: Random House, Koib. D. (1984). fiPeriendaf feOrning: Experience (IS the source of/ earning and deuelopmen). Engle.

wood Cliffs, NJ: Prentice- Hall. Koscbmano. T. (Ed.). (1992). Computer support for collaborative learning: Experience, theory

and design [Special 1ssueJ. ACM S/ ccuE ourlook, zf( 3). Koscbmaw T. (f994a). Toward a theory of computer- support for collaborative fearnk, g. ~o”,,,~ f

of Ihe Learning Sciences, 3, 218- 224. Koschmaw T. (1994b). The dialectics of instructional technology: Review of ~,,, r, pul~ m os m8

nitiue look. Educational Researcher, 23( 7), 38- 40. Koscbmaw T. (in Press). Logws- Latin redux: Review of The chi/ drenS machine. fnrerocdw

Learning Enoironments. ffischmaon+ T. 8 Newman. D.. Woodruff, E.. Pea. R.. & Rowley. P. (1993). Technology and pedagogy

forcoiiaboratfve problem solving as a context for learning: Report oo a cscw ‘92 workshop. ACM SKXHI Bulletin, ZS( 4). .57+ 41. Koschmaw T.. & O’Maiiey, C. (1994, October). Tutorial on compu( er sepporr for cd,,, bmdw

learning. CSCW ‘94 Conference, Chapel Hill, NC. Kuhn, T. S. (1972). The SfmCb~ re of sciendfii reoduJions (2nd ed.). Chicago: University r, l Chicago

PWSS. Kuuti. K. (1990 Activity theory as a potential framework for human- computer fnteraCtion re.

search. In B. Nerdi (Ed.). Confexf and consciousness.. Actiuity theory and humon cmpu, e, infemcdon (PP. 17- 44). Cambridge, MA: MIT Press, Lagemaoo. E. (1989). The Pioral worlds of educational research. History of Educorim ~~~ rrer, y,

29, 163- 214. Lalole. S.. & Dew, S. (Eds.). (1993). C~ mpurers M cognitiw mols. Hillsdale, NJ: ~~~ re,,~ e Erlbaum

Associates. Larkin. J..& Cbabay. R. (Eds.).( 1992). C

Hillsdale, NJ: Lawrence Erlbaum Associates. ~pu@ r~ ssisfed instmction and fntefkgent futoring systems. Law J.. & We%? er. E. (1991). Sihmted learning. Legitimate peripheml PQ&@,,&,. cambrfdge,

UK: Cambridge University Press. Lehnert. W. (i9&). Paradfgmatk kSues in cognitive science. In W. Kintsh, J. ~llie* & p. poison

(Eds.). Metho& and !actics in cagnifiue science (pp. 21- 49). Hillsdale, NJ: bweAce Erlbau,,, Associates. Lahrer. R. & Littlefield, J. (1993). Reiatlonships among cognitive components in bgo learning

and transfer. Journal of Educotionol Rychdogy. ss. 317- 330. Leoot’ev. A. N. (1974). The problem of activity in psychology. s~ ie,&~ hofw, 13, 4- 33. LePPer. f% Wooivertoo. M.. Mumme, D., & Gurtner, J. (1993). Motivational techniques of expert

human tutors: Lessons for the design of computer- based tutors. in S. bjoie & S. ear,,, (Eds.), CompurerS 0s cognitive rods (PP. 75- 105). Hillsdale. NJ: Lawrence Eribaom Associates, Mayer, R. E. (Ed.). (1988). T h‘

SPecrkw Hillsdale. NJ: Lawrence Erlbaum Asso& tes. eat lflg and fearning computer programming: Mu/ tip/ e research per. Merriii. M. D., Schneider. E. W.. & Fletcher. K. A. (1980). T/ CCI)-(/~ slrocriono/ DesiSln Lfbmry “of

40). Eogfewood Cliffs, NJ: Educational Technology Publicatfoos. Me** N, M. D.. Lin. 2.. & J ones, Twhndogv, 3q2). 7- 14. M. K. (1990). Second generation instructional design. Edu,- edono,

More. D.. & Ralph. C. L. (1992). A test of effectiveness of courseware in a college biology class. Journal of Educatmnol Technology Systems, 21. 7%~. Noddings. N. (1989). Theoretical and practical crmcerns about small groups io ,,, atbematics,

Elementary Schm( JOU~~ OI, 89. 607+ 3. Newman, D.. Griffin. P.. & Cole. M. (1989). Tke construction zone: Working for cqnreue ckong in

sChod. Cambridge, UK: Cambridge University press.

-- KOSCHMANN 12 ~M~,,~ Y, c. (Ed.). (1995). Computer supp& ed coJlaboratiue learning. Berlin: Sprl” ger- Verlag. @shea, T., &self, J. (1983). Learning and feaching with compulen. SLJSS~ X, UK: Ha~ estor Press. Pat,, mbo, D. B. (I~! xJ). Programming language/ problem solving research. Re~ ieU, ofEdaca/ fonaf

Research, 60. G- 89. Papert, S. (fg80). Minds( orms. New York: Basic Books. pea g,, & ~~, kd, 0. M. (1987). On the cognitive effeCtS Of learning Computer PrOgramming.

;” R, Pea & K. Sh? ingold (Eds.), Mirrors olminds: PalWns ofexperience in educadonaJmmputing cpp. 147- 177). Norwood. NJ: Ablex. piaget J, (,$ gj). T/ W e~ uilibraticm of cognihw structures: The central problem of infeffechwf de&

when,( T. Brown & K. J. Thampy, Trans.). Chicago: University of Chicago Press. Pyfyshyn, Z. (1989). Computing t” cognitive science. I” M. Posner (Ed.), Foundalions ofco@ oe

sc; ence (pp. 51- 91). Cambridge, MA: MIT Press. Riding, R., & Chambers, P. (1992). CDROM verses textbook: A comparison of the use of two

learning media by higher education students. Educational and Training TKhndW lnWn@ Wmf, 29.342- 349. Rosche~ te J., & Behro” d. S. (1995). The construction of shared knowledge in collaborative

probtkm solving. In C. O’Malley (Ed.), Computer- supporred cdJab0ratiW learning (Pp. 69- 97). Berlin: Springer- Vertag. Roth, M. (in press). The coevolutto” of situated language and physics knowing. JOUrnaf d

Science Educarion and Techndogy Rulko$., ska, J., & Crook, C. (Eds.).( 1987). Campurers. cognirion, anddeuelopmenl New York: Wiley. Salonron c. & Perkins, 0. N. (1987). Transfer of cognitive skills form programmf” g: When and

bow?‘ J~ rnd of& ducakond Computing Research, 3, 149- 169. Scandora J. & Scandura, A. (1988). A structured approach to intelligent tutoring. tn D. Jo” asSe”

(Ed.), &a& onaf designs for micrccompak’r course~ ore (pp. 347- 379). Hillsdale. NJ: Lawrence Erlbaum Associates. Scherz, Z., Goldberg, D., & Fund, Z. (1990). Cognitive implications Of learning PrOb!: Mistakes

and mkconcep~ tons, Journal of EducadonaJ Compuling Research. 6. 8!+ 110. Scbnase, J. L., & ~unnius, E. L. (Eds.). (1995). Proceedings of CSCL ‘95. Mahwah. NJ: Lawrence

Erlbaum Associates. S& ommer, M. (1990). Effects of b& efs about the nature of knowledge on comprehension.

.Jo,, rno, 01 Educarional Psychology. 82. 498- 504. &ott, T., Cole, M., & Engel, M. (19%). Computers and education: A cultural COnStrUCtiViSI per-

spective. Reuieu of Research in Educarion, 18, 191- 251. Segal, J. W., Chipma”, S.. & Glaser. R. (Ed%). (19%). Thinking and learning skills (Vol. 1). ReJadng

instrucrion ro research. Hillsdale. NJ: Lawrence Erlbaum Associates. Sham”, S. (~ Q~ XI). Cooperative learning in small groups: Recent methods and effects on achieve-

ment attitudes, and ethnic r& lions. Reuiew o/ Educational Research. 50. 241- 271. Shottw, ;. (1993). C~ JturaJ pditics of everyday life. Sccial conslruc~ ionism, rheforic and knowing of

Ihe lhird kind. Buckingham. UK: Open University Press. Shutman, L., & Keisler, E. R. (Ed?.). (1966). Learning by discovery. A critical appmisd. Chicago:

Rand McNally. Simon, H. (1969). Sciences of Ihe ortiliciol. Cambridge, MA: MIT Press. Simon, H. (1979). Information processing models of cognition. Annual Reoiew of Psychologx 30.

363- 396. Skinner, B. F. (1968). The techndogv olhmching. New York: Appleton- Century- Crofts. Soloway, E. (1993, April). Tutorial on mlerachve learning enuironmenls. InterCHl ‘93, Amsterdam.

Netherlands. Star S. L. (1996). Working together: Symbolic interactionism, activity theory, and information

6ystems. I” Y. Engestrdm & D. Middleton (Ed%). Communicah’on and cogmrion at work. Cam- bridge, UK: Cambridge University Press.

1. PARADIGM SHfFf’S AND INSTRUCTIONAL TECHNOLOGY 23

Steffe. L.. & Gale. J. (Eds.). (1995). ConstmctJuJsm in educadon. Hillsdale. NJ: Lawrence &jbaom Associates. St& r, F. (199.5). From unhwsing to conversing: A” ecological constructionist approach to

learning and multiple descrlptfo”. I” L. P. Steffe & J. Gale (Eds.), Constractiohm i” &wa& (pp. 6i-- 84). Htttsdale, NJ: Lawrence Erlbaum Associates. Steinberg, E. (1991). C~ PuJerfl~ isted JnsJrwian: A syn, hesis of theory. pmcdce, and JechnoJm.

Hillsdale. NJ: Lawrence Erlbaum Associates. suchm= n. J.. WW. Plans and situoled actions: The pmb/ em of human/ macb; ne c~ mun~~ d~.

Cambridge. UK: Cambridge University Press. SupPw P., & Macke”. E. (1978). The h& or& l path from research and development to opera-

tional use of CAL Educudonal Techndosy 18( 4). 9- 12. Taykw R. (Ed.). (lQfK% The mnpuler in the xhwC TuJor, wee, and I&. New York: Teacher

College ~‘* C. S. Thagard, P. (1992). CO~ C~ PIWIJ ~L& I~~ o” s. Princeton, N. I: Princeton University Press, van der Veer, R.. & Vatsiner. J. (1X- H). Understanding Vygmsky: A guest for symhesis. Oxford, UK:

Blackwell. VanLeh”. K. (1982). Bogs are “0t enough: Empirical studies of bugs. impasses and repairs f”

procedural skills. Journal of Mafhematicol Behaoior, 3, 3- 72. verzonl K. A.. & Swan, K. (1935). 0” the nature and development of conditional reaso” t” g t”

early adolescence. Applied Cogrtiliue Psychdw, 9. 213- 234. vo” Glasersfeld, E. (1979). Radical constructlvlsm and Piaget’s concept of knowledge. f” F. B.

Murray (Ed.). The impacr OfJ’iageJian theory( pp. 109- 122). Baltimore: University Park press. Vygotsky, L. S. (1978). Mind in sorie@ The dewlopmenr of higher psychdogical processes. can,.

bridge. MA: Harvard University Press. Webb. N. M. (1982). Student interactlo” and learning in small groups. Retie” of~ dwatiiona(

Research. 32. 421- 445 Wenger, E. (1967). Ard& iaJ in@ JJigence and Woringsyslems: Los Altos, CA: Morgan Kaufman. Wertsch. J.( fQ%). Vygolsk~ and fhesccialformafion dmind. Cambridge, MA: Harvard University

Press. Winograd. T., & Flares. F. (1986). Undersmnding computws and cqnirion. A nem (oundafion for

design. Norwood. NJ: Ablex. Woolf. B.. & McDonald. D. (13%). Contextdependent transitions in tutoring discourse. f” b

retdings of Ihe Nadonal Conference on Artificial lnlelligence (pp. 355- 361). ~0~ Altos, CA: nor. gan Kaufman. Wu. A. K. (199% December). A taxonomy of tutoring systems. I” T. W. Ghan (Ed.), Proceedings

&the 1993 JnWmWnaJ C~ nlerence M Computers in Educarion (pp. 187- 192). Taipei, Taiwan.