Technological Support for Knowledge Negotiation - Semantic Scholar

Proceedings of the 32nd Hawaii International Conference on System Sciences - 1999 Proceedings of the 32nd Hawaii International Conference on System Sciences - 1999

Technological Support for Knowledge Negotiation Joseph Petraglia Georgia Institute of Technology [email protected]

Bryce Glass Netscape Communications Corporation [email protected]

Abstract: As a result of a general dissatisfaction with transmission models of learning and theories of cognition that preclude context, "constructivism," or perhaps more accurately, a constructivist metatheory drawn from many sources, has come to dominate contemporary educational psychology. One of constructivism's principal implications, according to Sack, Soloway and Weingrad [9], is that "...activities of knowledge production (e.g., science) and reproduction (e.g., education) are about convincing, recruiting and enculturating others. in short a constructivist analysis of knowledge foregrounds rhetoric: the powers of persuasion and the difficulties of dispute" (p. 357). This rhetoricized conception of knowledge construction, we would argue, has not been adequately reflected in the work of educational technologists working from constructivist theory. Subtitled “A Rhetorical Approach to Constructivist Learning,” Reality Check is a prototype of educational software that we are currently developing to embody many of the lessons we believe knowledge negotiation offers educators and educational technologists. Using the metaphor of education as “knowledge-negotiation” Reality Check prompts educators and students independently to construct their representation of a task by identifying relevant variables and giving reasons and evidence that permit fruitful interrogation. By setting up alternative representations, Reality Check explicitly invites the teacher and student to participate in a negotiation of knowledge in which evidence is exchanged and positions are clarified, whereby each party in this educational activity can better understand the other and, more importantly, make productive use of this understanding.

1. Introduction Education has been affected by many shifts in other disciplines this century. Contemporary theories of language, learning, social behavior and knowledge have drawn back from what is generally characterized as an unwarranted faith in objectivity and the “self-evidence” of k n o w l e d g e . The model of education that these assumptions supported (and the model most educational theorists have rejected) is often called the “transmission” model of learning. According to this model, the principle challenges faced by teachers are essentially those of greasing the student's mental machinery and packaging discrete bits of information in such a way that learners can easily grasp and commit them to memory. Because knowledge is “out there,” according to the transmission model, the trick is to design pedagogies that efficiently relay that objective reality to the waiting minds of passive learners. For this reason, the transmission model of learning is wedded to the proposition that social interaction is separable from knowledge-building and that context and culture are immaterial. In contrast to the transmission model of education, many in education have adopted what might loosely be termed a “constructivist” approach to learning. Better yet, one might say we have adopted a constructivist metatheory as the approach is not cut whole-cloth from any single theory or theorist but is often an amalgam of up-dated Deweyan progressivism, Vygotskian sociohistoricism, and Piagetian ontogeneticism that centers on common themes such as the centrality of experience to learning, student-centeredness, and thinking

that is situated in “authentic” contexts. Although one must be careful not to erase the considerable differences between theorists such as Piaget and Vygotsky, most contemporary educators share with them and with Dewey the premises that we construct our knowledge of the world based on prior knowledge and experience, and that knowledge and learning are derived from participation in activities that are distributed across social, cultural and material spaces. Resnick argues [8, p. 2] that a constructivist position obliges us to view social behavior not just as an influence upon thought, but as itself a manifestation of cognitive processing which leads us to "analyze the ways in which people jointly construct knowledge under particular conditions of social purpose and interaction." Furthermore, as Dunn [4] has concluded, "if we do indeed construct our reality and learn nothing that is not somehow context-specific, then instruction should take place in rich contexts that reflect the real world and are as closely related as possible to contexts in which this knowledge would subsequently be used...in a word, they need to be authentic" (p. 84). A lack of authenticity is cited as a major obstacle faced by educators in many fields, but especially those educators in the sciences. In the words of Burbules and Linn [2] Students rarely see a relationship between the science they learn in school and the science problems they encounter in everyday life. This narrowness is attributable not only to the generally recognized difficulty of transferring knowledge from one domain to another, but also to an active belief on the

0-7695-0001-3/99 $10.00 (c) 1999 IEEE

1


part of students that ‘school knowledge’ represents a distinct and special category of learning, separate from the commonsense solutions they develop in real-life contexts” (228). The integration of school with the world outside school is thus seen as key to educational success. Constructivism's mandate for authenticating learning is clearly observed in the field of educational technology where even basic technological dimensions and modalities such as interactivity, networking, and multimedia are reenvisaged and recast as features that further the constructivist goal of making school learning correspond, to as great a degree possible, to the learning that results from everyday interaction [Hmelo & Nararyanan, 5]. There is a problem, however, in that for educational technologists, authenticity has been subject to often incomplete operationalizations and equations: authenticity lies in factual accuracy (e.g., as in goal-based scenarios), or in a problem’s complexity (e.g., anchored instruction), or in collaborative activity (a component of many environments), strategic mentoring (e.g., cognitive apprenticeship), and/or multiple representations of the task (e.g., cognitive flexibility) etc. In each of these instances, the authentic nature of the task depends on a feature that can be satisfactorily verified through external and independent means. Given such operationalizations, it seems safe to presume that one can, prior to entering a given teaching situation, make a problem or a learning environment authentic. in a word, it is generally assumed that educators can preauthenticate. In making this assumption, educators takes on certain beliefs about the nature of the real world and the possibilities of knowing. For instance, when someone labels something "authentic" he or she is making not only a statement about that thing, but about the world of which that thing is an exemplar. The declaration of a biology problem's "authenticity” for instance, tells us more about the speaker's notion of what “real” biology is like than it does about the problem being described. In the case of the Jasper Woodbury anchored instruction materials, for instance, or goal-based scenarios such as Broadcast News, developers presume that because they have provided factual information in informationally dense multi-mediated environments, that they have ensured the task's authenticity. But, if one accepts the constructivist tenet that cognition cannot be divorced from the contexts in which it transpires, we must also accept that an authentic assignment cannot be manufactured independent of the learner’s assessment. In other words, authenticity, like beauty, is in the eye of the beholder and thus authenticity is a judgment rather than a quality that inheres in a task. In essence, then, preauthentication contradicts the studentcenteredness that is the very soul of constructivism. The dilemma preauthentication points to is a very real one as it suggests that the bedrock upon which our entire

theory of learning rests is put in jeopardy. If we find fault with preauthentication (as constructivists should) our options for engaging students in real-world learning may seem severely constrained. In response to this threat, we propose that an approach which some are calling Knowledge Negotiation (KN) provides a framework that helps resolve the challenge preauthentication poses to constructivist educators by explicitly emphasizing the dialogic nature of learning and accentuating the role of negotiation and persuasion.

2. Knowledge Negotiation 2.1. The KN framework While lauding the opportunities technology affords constructivism, Vosniadou [10, p. 13] acknowledges that constructivism's application is not unproblematic, and that "Students must still realize the similarities and differences between the simulations to which they have been exposed and in the multi-media environments and the real life tasks." Teachers, she argues, must be able to articulate those differences and lend support for the similarities. in this, she is seconded by Driver et al. [3] who argue that “if students are to adopt scientific ways of knowing, then intervention and negotiation with an authority, usually the teacher, is essential. Here, the critical feature is the nature of the dialogic process” (p. 3). As Driver and her colleagues make clear, the challenge educators face has less to do with giving out the right information, than in engaging the learner in a process of negotiation that provides an opportunity to articulate his or her representation of an assignment and thereby offers the educator the best hope of intervening in that representation. As Lakoff and Johnson [5] note: When people who are talking don't share the same culture, knowledge, values and assumptions, mutual understanding can be especially difficult. Such understanding is possible through the negotiation of meaning. To negotiate with someone, you have to become aware of, and respect, both the differences in your backgrounds and when these differences are important. (231) In Europe (and especially in Britain) computer scientists working on tutoring systems have produced a distinctive approach that, we believe, addresses constructivist learning in a way that acknowledges its dialogic dimension. Moyse and Elsom-Cook [6] have dubbed this approach "knowledge negotiation." Unlike the more typical environments that contain the correct knowledge and are to be used to lead students to the ideal understanding the environment already possesses (the "full technology" approach to educational technology), a knowledge negotiating system does not presume to possess the all the ansirs and instead seeks to negotiate problem solutions with learners. Although some of its

0-7695-0001-3/99 $10.00 (c) 1999 IEEE

2


tenets have precursors in other intelligent tutoring systems, Moyse and Elsom-Cook correctly argue that this approach is unique in that earlier learning environments that were designed to permit students to explore a variety of viewpoints (e.g., Papert’s Logo) only permitted this exploration in predetermined ways. Like cognitive flexibility theory, the KN approach holds that there is a need to represent multiple viewpoints, but unlike the cognitive flexibility framework, KN explicitly accepts that in many instances there may not be any single pre-existing truth to be conveyed, only a warranted belief. Baker [1] has summarized the KN approach to intelligent tutoring by listing eight of its tenets: 1. Some (or all) tutors do not possess complete knowledge for the domains which they teach. 2. For some domains there is no single, privileged or correct viewpoint on knowledge. 3. There exists domain, or parts of them, for which there is no "knowledge," but only a number of different competing and arguable sets of justifiable beliefs. 4. For some problems in some domains there are multiple correct or acceptable solutions. 5. For domains that possess some of all or characteristics 1 to 4, teachers should not aim to simply transmit their own knowledge (beliefs, viewpoints, solutions, etc.) to the student. 6. The tutor should not have complete control over the tutor-learner interaction in domains that possess some or all of characteristics 1 to 4 above. 7 . The representation of the domain to be taught/learned, and the problem solutions themselves, should be jointly constructed by teacher and learner. 8. Negotiation mechanisms and representations to support them are required in teaching/learning interactions with respect to domains that possess some or all of the properties 1 to 4. By framing learning as a process of negotiation, a KN approach directs our attention towards the gathering of information and evidence, the arrangement of informational inputs in some readily accessible way, the application of reason to the construction of claims and to the association and disassociation of different claims, and to the exercise of judgment concerning what action to take. Such framing makes constructive processes less abstract and metaphorical and more amenable to useful intervention, for unless the teacher can figure out how (and why) students arrive at the ansirs or solutions they do, there is little chance that he or she can productively engage or modify their conceptual structures. Burbules and Linn [2] have approached essentially the same interests in negotiation from a slightly different angle. They argue that their approach is meant to interest and motivate students who currently are uninvolved in science learning; to help them see the relevance of science learning to their life outside of school; and to take

advantage of their social interests to facilitate better science learning. This approach will also give students a more authentic understanding of how science actually happens. Science as a human activity is not pristine, impersonal or abstract, nor does it proceed by a single method. Science education should reflect this reality. (239) Negotiation is also a reality of science. Science is accomplished through complex social interaction and argumentation over the adequacy of opertionalizations, whether evidence is sufficient to make a particular claim, and ultimately, the correct interpretation of data. In this sense, knowledge negotiation in the classroom mirrors how science -- as ill as social sciences and humanities -are undertaken in the real world. In the words of Sack, Soloway and Weingrad, "...activities of knowledge production (e.g., science) and reproduction (e.g., education) are about convincing, recruiting and enculturating others. In short, a constructivist analysis of knowledge foregrounds rhetoric: the poirs of persuasion and the difficulties of dispute" [9, p. 357]. Framing education as a sort of knowledge negotiation opens interesting theoretical questions: How do prior assumptions take the shape of claims? What counts as “evidence” to students and why? Do different sciences have different argumentative structures, and if so, how do they vary? How do experts, as opposed to novices, argue for their understanding? Within the scope of the classroom, KN has a special significance and negotiation serves special purposes. It: • Makes prior assumptions (one's own and others') visible, • Presents opportunities for clarification and refinement, • Reveals science's foundations in argument and interpretation, • Helps teachers tailor instruction to students, • Promotes critical thinking, • Highlights the role of communication in learning, and • Puts the teacher-student relationship at the center of the classroom.

2.2. How radical a framework is it? In what is sometimes called the postpositive paradigm, Burbules and Linn argue, constructivist educators must walk a line between two inadequate views of science which called be labeled “positivistic” and “relativistic.” The positivistic approach essentially proposes that the external world is a fixed reality can be described as it really is. Thus, science (when done correctly) produces truths that are above contestation, according to this view. The relativistic approach suggests the opposite: that because we have no indisputable foundations from which to view the world, no knowledge

0-7695-0001-3/99 $10.00 (c) 1999 IEEE

3


claim is any more defensible than another. Neither of these views offers much hope to the science educator, they argue because neither reflects scientific activity as it is actually performed. We would argue that KN (indeed, constructivism generally) successfully avoids being entrapped in either positivism or relativism. Of course, a KN approach to learning has to overcome some cultural obstacles, for being an educator has traditionally meant foregoing belief in favor of “objective” truth. But a KN approach merely rearticulates the point, made by various constructivist educators and psychologists over the years, that the idiosyncratic construction of problems and their solutions by learners is unavoidable. The manner in which a student's problem-solving process unfolds--even when grossly inaccurate, inefficient, or otherwise unacceptable from an educator's perspective--is not random, but reflects prior efforts on the student’s part to make sense of the task in a way that conforms with his or her knowledge of the world. For the truly constructivist educator, then, the idea that we negotiate knowledge is a statement of fact rather than a recommendation. We wish to be very clear that our use of the term "negotiation" does not suggest that the knowledge bazaar is open, nor does it imply that teachers should relinquish authority over classroom interaction. On the contrary, the KN approach places the teacher in the position of an expert who understands the limits of expertise and who appreciates that students may have important understandings of tasks that require productive engagement if the teacher's lessons are to take hold. For many of the same reasons that naturalistic research methods have relied on participant observation, a knowledge negotiating perspective on education asks teachers to accept the importance of their own assumptions and meaning-making practices, for as tasksetters, informational resources and arbiters of performance, teacher cognition has an enormous impact upon student learning. While we appreciate and approve of the spirit in which student-centered approaches are offered, we feel (as most educators do) that the classroom is where teachers and students must actively work together.

3. Technological Support for Knowledge Negotiation: Reality Check Writing to an audience of educational technologists, Vosniadou [10] best articulates, we believe, the nature of the practical challenges facing educators generally. The creation of learning environments that allow existing representations to be externalized and examined can facilitate conceptual change and can contribute to the creation of metaconceptual awareness. . . Such learning environments help students to make their representations public and to examine them

objectively. It makes it possible for students to draw the hidden implications of these representations, to discuss them with others, and to change them (15). Subtitled “A Rhetorical Approach to Constructivist Learning,” Reality Check is a prototype of educational software that we and another colleague are currently developing that embodies many of the lessons we believe knowledge negotiation offers to the constructivist educator and to the educational technologist.1 Using the metaphor of education-as-argumentation, Reality Check prompts educators and students independently to construct a task “profile” by identifying relevant variables and giving reasons and evidence for their relevance in order to articulate and compare their senses of how elements of the task correspond both to the assignment and to the “real world.” Reality Check does not support construction of a neatly rational profile of the task from either the educator’s or the learner’s perspectives nor does it attempt to force closure upon points of discrepancy, but instead tries to make explicit and describe as fully and multidimensionally as possible, the gap between the two. In keeping with the guidelines outlined elsewhere in this paper, Reality Check is only intended for use in those circumstances where the disciplinary content, the educator's objectives, and the learner's intellectual development are appropriate. This means that it is not intended for use with ill-structured problems where transmission, rather than negotiation, may be warranted, nor with tasks where the teacher finds pedagogical value in rigidly stipulating a context within which they must be performed. The four steps that make up Reality Check’s protocol are 1) the entry of the assignment 2) a teacher profiling activity, 3) a student profiling activity, and finally, 4) a joint teacher-student analysis. In the first step, the teacher simply pastes a text file of an assignment into Reality Check. After the assignment has been entered, Reality Check then prompts the teacher to identify three criteria he or she wishes to use for "profiling" the assignment. Examples of such criteria might be "parts of the task requiring greatest amount of student creativity and initiative," or "parts of the task that will figure most heavily in its evaluation" (although the prototype currently provides a menu of criteria, it also accommodates the teacher's desire to create his or her own criteria). Once the teacher has selected the criteria, Reality Check prompts the teacher to mark up the assignment by highlighting a limited number of sentences, phrases or words of the assignment that he or she judges to be most associated with each of the three criterion. Reality Check then leads the teacher to go back and annotate the selected

1 Reality Check is part of a NewMedia Center project funded, in part, by the EduTech Institute and the Graphics, Visualization, and Usability Center at Georgia Tech. In addition to ourselves, the design team includes Greg VanHoosier-Carey of Georgia Tech.

0-7695-0001-3/99 $10.00 (c) 1999 IEEE

4


elements of the assignment with justifications for why the various choices were made. Subsequent to the teacher’s highlighting and annotation, students then independently mark up the assignment much as the teacher did using the criteria determined by the teacher. Just as their teacher did, students are then returned to the result of each mark-up routine and asked to annotate the elements to justify their selection and prioritization of elements along each criterion. Once both the teacher and the student have profiled their understandings of the assignment, Reality Check uses the highlightings and annotations to contrast the variables and the evidence each side finds relevant both in performing the task as part of an academic assignment as well as what each finds relevant in a real-world execution of the task. It also places the completed profiles side-byside and suggests areas where the educator and learner appear to have serious disagreements. For instance, Reality Check might flag places where the teacher thinks it is important that students use tools he or she normally associates with the task, whereas students may not realize these tools exist, or that they are permitted to use them, or that the employment of these tools will really make much of a difference. Or perhaps Reality Check would signal that the student believes that the designated evaluator of the task makes little difference to the tasks execution whereas the teacher thinks this is a critical consideration and will be looking for sensitivity to this issue in his or her assessment of the student’s performance. Thus, the desired outcome of a teacher-student dyad’s use of Reality Check is a textual as well as a graphical representation of the task that identifies the variables and the evidence each "side" finds relevant in performing the task. In the final and most important step, the teacher and student come together to discuss these representions and to negotiate their divergent understandings. In this negotiating phase, teachers and students may discover that they are in more agreement than Reality Check's analysis suggested, for Reality Check's purpose is not so much to produce an accurate picture of the gap between the student and teacher's task representation but rather to provide an opportunity for negotiation of the task that might not happen otherwise. In other words, Reality Check's real function is to prompt sustained reflection of one's task representation in relation to another's and to provide a framework within which teachers and students can carry on a dialogue. Although this is a necessarily sketchy overview of Reality Check, it may be sufficient to illustrate several of the ways in which on might design a technology that supports knowledge negotiation objectives. First, a system such as Reality Check demonstrates how, instead of presenting learners with preauthenticated materials as is currently done, technology can be used to challenge the learner to reflect on the contextual disjunctures they

encounter in the environment and to speculate on the implications these disjunctures might have for their performance of the task. This does several things, but perhaps most importantly (to use Nix's [7] terminology) it dignifies the learner by making explicit the experiential superiority of the human learner to the mechanistic environment. While lending learners dignity might be a worthwhile end in itself, as educators we might find dignity useful; it makes the learner cognizant of the importance and relevance of his or her own prior knowledge and understanding and in doing so, it makes the learner an active participant in the learning experience. By inviting learner rejoinders to the educator's construction of a problem space, we shift the didactic nature of instruction more towards genuine coconstruction. Engaging the learner in this sort of dialogue may give learners a metacognitive awareness of their own reasoning and invites students to consider the ways in which their task representations differ from their teachers’ as well as from their peers. From a rhetorical perspective, one of the appealing attributes of a technological support such as Reality Check is that it explicitly acknowledges -- indeed, emphasizes -- the school-boundedness of the assignment's performance. As mundane an observation as this is, perhaps, the separation of the context of schooling from the contexts in which knowledge may come in useful outside of school is one that attempts at preauthentication tend to mask. It seems natural that the educator might find value in having students explain how their understanding of the assignment differs from the his or her own. This provides the educator with a wealth of information about how learners are interpreting the assignment and constructing the problem space and may assist the teacher in discovering ways of overcoming cognitive difficulties students encounter. Furthermore, if patterns of learner reactions seem to point to similar disjunctures between their understanding of the problem and the teacher's representation of the problem, this would provide a valuable check on the appropriateness of the problem's contextualization for the group of learners it is meant to support. Finally, and most fundamentally, perhaps, a system such as Reality Check highlights the negotiative nature of education and allows the teacher to present his or her case for the importance of the task to the overall assignment and the relevance of the assignment to the learner’s real world. By setting up alternative narratives in this way, Reality Check explicitly invites the teacher and student to participate in a kind of exchange in which evidence is articulated and positions are clarified, whereby each party in this educational activity can better understand the other and, more importantly, make productive use of this understanding. Reality Check thus pulls the educator into the knowledge-construction process in a way that permits students to see the educator

0-7695-0001-3/99 $10.00 (c) 1999 IEEE

5


as a reasonable, pedagogically-motivated individual rather than an arbitrary task-setter. Again, we believe that constructivism teaches that it is not educators' prerogative to permit alternative constructions of the way the world works. Instead, we must first acknowledge them as the logical consequence of a constructivist view of learning and then to engage these constructions in a manner we find beneficial to the learner. While a prototype such as Reality Check is hardly paradigmatic of a knowledge negotiating approach, we believe it nonetheless illustrates one way in which technologies can be used to highlight the role of argument in learning and thus realize the lessons of constructivism for education. References [1] Baker, M. (1994). A model for negotiation in teachinglearning dialogues. Journal of Artificial Intelligence in Education, 5 (2), 199-254. [2] Burbules, N. C., Linn, M. C. (1991). Science education and philosophy of science: congruence or contradiction? International Journal of Science Education. 13 (3), 227-241. [3] Driver, R., Asoko, H., Leach, J., Mortimer, E. & Scott, P. (1994). Constructing scientific knowledge in the classroom. Educational Researcher, 23 (7) 5-12. [4] Dunn, T. G. (1994). If we can’t contextualize it, should we teach it? Educational Technology Research and Development, 42 (3), 83-92. [5] Lakoff, G. & Johnson, M. (1980). Metaphors we Live by. Chicago: University of Chicago Press. [6] Moyse, R. & Elsom-Cook, M. (1992). K n o w l e d g e Negotiation. London: Academic Press. [7] Nix, D. (1990). Introduction. In D. Nix & R. Spiro (Eds.) Cognition, Education, and Multimedia: Exploring Ideas in High Technology. Hillsdale NJ: Lawrence Erlbaum. [8] Resnick, L. (1991). Shared cognition: Thinking as social practice. In L. Resnick, J. Levine & S. Teasley (eds.) Perspectives on Socially Shared Cognition. American Psychological Association: Washington DC. 1-20. [9] Sack, W. Soloway, E. & Weingrad, P. (1992) Re-writing cartesian student models. Journal of Artificial Intelligence in Education, 3 (4) 381-402. [10] Vosniadou, S. (1994) From cognitive theory to educational technology. In S. Vosniadou, E. De Corte and H. Mandl (Eds.) Technology-based Learning Environments. Berlin: Springer. 11-17.

0-7695-0001-3/99 $10.00 (c) 1999 IEEE

6