7.4 Reexamining Some Key Concepts

In this section we would like to review some of the key concepts in instructional design and instructional methods, examining them from our constructivist perspective. In this discussion, it should be clear that methods can be implemented in many different ways, and how a method is implemented and what is the focus in that implementation is reflective of one's views of learning.

7.4.1 Discovery Learning

Discovery learning has a long and complex history in education (see Dewey, 1929; Bruner, 1961; Page, 1990). While it reached its heyday as a pedagogical framework in the 1960s, the generality of the term allows it to be applied to any learning environment in which the student is actively involved in problem solving (Bruner, 196 1).

Discovery learning in its original formulation focused on the learning process, the goal being to develop inquiry skills in a content domain, an appreciation of inquiry as a way of approaching issues, and an appreciation of the complex issues in a domain. This view of defining what is learned in terms of the interrelationship of process and content is perhaps best exemplified in Bruner's design of the social studies curriculum, MACOS (Bruner, 1966; Bruner & Dow, undated). However, in its implementation over the years, in the form of the "new math," "open classroom," and other movements, discovery learning was employed as a method for acquiring content. The content goals remained the same as for other learning environments--knowledge was still seen as an entity--but the strategy for acquisition changed to one of "discovery."

The consequence of this view of discovery is that the students' inquiry is not honored. Rather, the learner has to discover the answer that the teacher already knows. Needless to say, learners quickly discover that the goal is not inquiry or exploration of a domain but rather discovering what the teacher wants them to discover. Rather than learner centered, the instruction is quite clearly teacher centered.

This view of discovery is perhaps an unfortunate consequence of the metaphor; "to discover" suggests that there is something (knowledge) hidden away, and our job is to find and acquire it. An alternative view of discovery is to think of it in terms of "invention," a personal construction, rather than as a discovery of what exists. From this perspective, we take as the goal of instruction not the acquisition of a specific, well-defined bit of content but rather the ability to learn in a content domain. Learning to learn--including the ability to ask questions, evaluate one's strategies, and develop answers to questions in the content domain--is the goal in this view of discovery learning (Brown et al., 1993). Such a goal requires a unity of process and content; both are integral and inseparable in developing the ability to work and think in the content domain. In this learning-to-learn view of discovery learning, the knowledge is in the learner's activity rather than being in the text. Thus it is a view that is fully consistent with the constructivist viewpoints.

In summary, if the goal is simply to learn a well-defined content--definition and procedures--then a discovery approach is not necessary. The learner should simply be told the answers and either given a memory (Job) aid or required to memorize it. However, if the goal is to be able to use the information in a content domain, to be able to think in the content domain, to be able to invent defensible understandings, then the discovery method is appropriate.

7.4.2 Zone of Proximal Development

Vygotsky (1978, p. 86) defines the zone of proximal development (Zo-ped) as "the distance between the actual developmental level of a child as determined by independent problem solving and the level of potential development as determined through problem solving under adult guidance or in collaboration with more capable peers." In other words, we are describing a form of "joint" cognition, where the tutor provides support or scaffolding for the individual until the individual appropriates the knowledge or skill and brings it under his conscious control for his own use. The support is progressively withdrawn and, as the students take over more and more responsibility in a problem -solving situation, they become self-regulated and independent.

We would broaden the focus in two interrelated ways. First, there is a matter of perspective. Rather than talking about what we do to an individual, we would prefer to discuss the affordances of the environment. Thus, we can look more broadly at the environment to determine how the environment is designed to be supportive of the individual in relation to accomplishing some task. Neither the student nor the teacher "owns" the Zo-ped; rather, it is something that is established dynamically. From an instructional design perspective, this shifts the focus from what we teach to how we design a learning environment that will support the learner as he or she may request support, and can be discarded by that learner when it is no longer required.

Second, our broader view of the Zo-ped looks at the full cultural context of the individual's learning environment. Cole (1985) has characterized the Zo-ped as "where culture and cognition create each other." We can think of the Zo-ped in terms of what an individual can do as a function of being a part of. a specific culture that would not be possible if he or she were not a part of that culture, and, as Cole's remark suggests, the changes in the individual, in turn, change the culture. Thus the full sociohistorical context is a part of the Zo-ped for development. An implication of this is that we can look at what a Zo-ped affords in any cultural context by simply looking at the difference between newcomers and old-timers in that community (Lave & Wenger, 1991).

This view has interesting implications for the distinction between learning and instruction (see, e.g., Heinich, Molenda & Russell, 1993) where instruction is defined as particular context for learning in which we purposefully organize the environment to achieve particular learning objectives. In the view we are discussing, there are a wide range of social situations that are designed to promote learning of particular goals. In our view, these social contexts are instructional environments. They differ from what we normally describe as instruction simply because we typically conceive of instruction as the formalized delivery or transmission of information. However, there are designed formal and informal social structures that teach the newcomer the rules and procedures of conduct. For example, Lave and Wenger (1991) describe the design of. AA meetings such that individuals stand up and share stories. They provide a very nice analysis of this "instructional strategy" and its impact on newcomers. More generally, we could examine any community and look at the structure of the community as well as the effect that structure has on how we learn to behave. Some instructional designers might want to argue that such an environment is not designed, and hence it is not instructional. However, we would argue very strongly that it is designed to achieve particular ends. Simply try to change the structure of the community (the tacit rules) and see how the members would react. And of course, anyone who fails to learn is redirected in a manner analogous to any training program.

7.4.3 Scaffolding

Success in the Zo-ped requires support for learning, and that support is called the scaffolding. Indeed, the Zo-ped is defined in terms of the scaffolding or affordances of the environment. Scaffolding includes the support of other individuals, any artifacts in the environment that afford support, as well as the cultural context and history the individuals bring to the Zo-ped. We will discuss the role of the teacher in greater detail in the next two sections. At this point, we simply wish to clarify the general notion of the power relationship within the Zo-ped.

The scaffolding metaphor implies a rigid structure that is used to construct. That is, the placing of the scaffold presupposes much of the character of the structure being built. In our mind, this is an unfortunate choice of metaphors since it suggests a guiding and teaching of the learner toward some well-defined (structural) end. This "structural" metaphor of a scaffold is consistent with the objectivist view of instruction in which the teacher arranges the environment (including the teaching activities) to help the learner acquire the prespecified "knowledge." This view 0f scaffolding presents the Zo-ped as a teaching environment rather than as a learning environment. It is a model of transmission, the "objects" that provide the support/scaffold form the individual.

From our perspective, the Zo-ped and the scaffolding must be viewed as a learning environment--as supporting the growth of the learner. Griffin and Cole provide what we feel is an excellent contrast of the learning vs. teaching concept of scaffolding in the Zo-ped: "a Zo-ped is a dialogue between the child and his future; it is not a dialogue between the child and the adult's past" (1984, p. 62).

7.4.4 Cognitive Apprenticeship

The influential papers of Resnick (1987) and Brown, Collins, and Duiguid (1989) led to a renewal of interest in apprenticeship as a design for learning environments. The focus, however, shifted from physical job skills to the development of cognitive skills. The result was a focus on authentic learning environments where the cognitive demands in the learning are qualitatively 'the same as the cognitive demands of the environment for which the instruction was preparatory. We fully agree with the focus on authentic cognitive demands. Indeed, this is consistent with our emphasis that the learning is in the activity of the learner, and hence we must examine the activity and the full sociocultural context in which it occurs. In traditional instruction, the learner's cognitive activity is centered on the development of strategies

for determining what the text and the teacher are signaling as important, processing and remembering the information, and for evaluating test items to determine correct answers. These are all skills preparatory for more schooling--it is a cognitive apprenticeship for schooling--but not for much else (Honebein, Duffy & Fishman, 1993). Engaging learners in cognitive and metacognitive activities that involve the authentic use of information is a central goal in our instructional design.

While we agree with the focus on authentic cognitive activity, there are other aspects in the development of the concept of cognitive apprenticeship with which we are less sanguine. Our primary concern is that there has been a focus on cognitive apprenticeship as a "master-apprentice" relationship with an implicit view that the core of the apprenticeship is the master teaching the apprentice. Indeed, the most often-cited examples of cognitive apprenticeship are reciprocal teaching (Palinscar & Brown, 1984) and modeling "thinking like a mathematician" (Schoenfeld, 1991), both of which focus almost exclusively on a knowledgeable master working with the less-knowledgeable learner. This model of cognitive apprenticeship is more in line with the MAB or even with the MAC models of mind. It focuses on the individual cognitive activity (Kang, 1995).

Lave and Wenger (1991) present a view of cognitive apprenticeship more in line with our MAR model of mind. Rather than the master-apprentice relationship being central, Lave and Wenger examine the full sociocultural context of which the individual is a part. They discuss apprenticeship as being a legitimate peripheral participant in a social context. By this they mean that the individual is legitimately a participant but is only playing a partial role in the context. There is not a master who assigns tasks or who monitors the apprentice's behavior. Rather, the apprentice begins to assume responsibilities, testing his or her ability to assume roles and responsibilities in that environment. The full cultural context--the artifacts as well as the experts (or old-timers)--afford the learner support (scaffold) as the learner attempts to take on these responsibilities. Lave and Wenger note that in most apprentice environments there is little direct teaching between master and apprentice.

In addition to decentering the "master" in the apprenticeship environment, Lave and Wenger argue that it is not so much that apprenticeship is the "best" learning environment; it is simply that it is a prevalent learning environment. In addition to formal apprenticeship, most informal learning can be interpreted from the perspective of "apprenticing" (see 20.3. 1). In most new contexts, we first observe and then begin to take on some responsibilities in a group we wish to become an integral part of. Lave and Wenger provide several excellent examples of both formal and informal apprenticeships as a means to illustrate their concept of legitimate peripheral participation and to begin to analyze some of the critical variables in successful apprenticeships.

In summary, from an instructional design perspective, the apprentice environment is one way to view the design of a learning environment. However, in doing so the emphasis is not on master-apprentice but rather on the learner as a member of a larger community of practice who, through legitimate peripheral participation and the affordances of the environment, begins to assume greater responsibility in that community of practice. Thus our design must provide the learner access to that community of practice and provide the tools that will support the learner in assuming his or her role in that practice. The instructional principles outlined in the previous section provide the guidance in designing such an environment.

7.4.5 Coaching

We no longer teach, but rather we coach--we have moved from the sage on the stage to the guide on the side. The coach provides the scaffold for the learner. This is becoming common rhetoric in instructional theory, and the constructivist "movement" has been a primary stimulus for this shift in the teacher/trainer role. It is not so much that the teacher is seen as less important, rather the role of the teacher changes so that the focus is on aiding or providing the scaffolding for the learners rather than telling the learner. We fully appreciate this goal of decentering the teacher as the fount of knowledge. However, our concern is that the shift is a shift in method rather than a shift in the conceptual framework underlying the method--a trivial rather than a radical constructivist shift in von Glasersfeld's (1989) terms.

By a shift in method, we mean that while the method has moved from sage on the stage to guide on the side, the guide is still the fount of knowledge. He or she still possesses the knowledge the student is to acquire. It is a unidirectional relationship in which the student observes and mimics or follows the instructions of the coach. The coach, in turn, models the behavior or provides the answers. We coach the learner by giving the learner our knowledge, which is to be replicated. This is in large part consistent with the cognitive apprentice model advocated by Brown et al. (1989). Of course, this view is not much different from the traditional view of learning. Knowledge is still this entity to be transmitted from coach to learner--a new instructional method, but the same view of learning and knowledge.

Within our MAR constructivist framework, the coach-learner relationship is bidirectional. The skills and knowledge of both coach and learner are attended to and honored. Fosnot (1989) describes this as a mentor-protégé relationship in which the mentor begins by seeking to understand and expand the learner's or protégé's current conceptions. Thus both mentor and protégé are seeking to understand the other's views. Rather than the end goal being that the learner can replicate the coach's behavior or follow the coach's directions, it is that the mentor and learner come to an agreement even if it is an agreement to disagree. That is, the learner may not mimic the coach, but the deviations are knowledgeable deviations that the learner can defend and the coach can respect.

We can see for ourselves, and hopefully illustrate for the reader, the implications of these two points of view in our interactions with students and, oftentimes, even in our interactions with each other as we discuss research and theory. The most "natural" approach to an issue-based discussion is to have generated one's own point of view, one's own answer to the question on the table. Then, in conversing (coaching) student or colleague, we listen to see how what they say matches our conceptualization, and we catalogue the matches and the mismatches. Our response, then, is a question or comment that helps the learner understand our different points of view and encourages them to accept the alternative. This is the coaching framework that calls for the learner or colleague to mimic our point of view.

The alternative framework--the mentoring approach--would have us listening to the student or colleague to understand his or her point of view. We would then ask questions to help us clarify those aspects that we did not understand. Only after we felt a reasonable understanding of that alternative view would we engage in a discussion to try to understand and perhaps resolve the differences. But it may well be that the alternatives are compatible and each is acceptable. Fosnot has offered the following prescription for coaching preservice teachers within the constructivist framework, a prescription that is a method reflective of the underlying theory. It should be noted that while it is similar to the Socratic method, the acknowledgment of defensible alternative perspectives provides the deviation from the mimic aspect of the Socratic method:

1. The mentor learns the protégé point of view through careful listening and probing.
2. The mentor teaches by inquiring at the "leading edge" of the protégé's thinking and by attempting to facilitate disequilibrium.
3. The mentor constructs a line of inquiry meaningful to the protégé , and the protégé constructs a line of reasoning meaningful to the mentor.
4. The mentor acknowledges that the protégé has the intellectual freedom to adopt and modify the pedagogical orientation of his or her choice (Fosnot, 1989, p. 97).

7.4.6 Context

In instructional and educational research, we have traditionally viewed context as a variable in our research. Context can be decomposed into components, and those components can be manipulated. The context sits separate from the individual and can be manipulated independently of the individual. This is an objectivist view of context more fitting of the MAC model of mind (see footnote 4, p. 172).

The constructivist view of context we would argue for has the context as a dynamic whole, including the individual and the sociohistorical context. The mind as rhizome provides one metaphor for this view, where elements can be pointed to much as the tubers, but in the context of the whole we cannot identify, where one element leaves off and another begins. Birdwhistell offers an analogous metaphor to explain this view of context:

... sometimes I like to think of a rope. The fibers that make up the rope are discontinuous; when you twist them together, you do not make them continuous, you make the thread continuous ... even though it may look in a thread as though each of those particles are going all through it, that isn't the case ... that's essentially the descriptive model (Birdwhistell as cited in McDermott, 1980).

7.4.7 Learner Control

Learner control is a concept that was introduced in relation to computer-based instruction (see 23.9). At issue is how much or what type of control should be given to the learner during the learning process? The alternatives are learner control, computer control, or shared control. What is it that is controlled? Basically the control decision has to do with the pacing of the information presented, the sequence of the information, and the actual content (Milheim. & Martin, 1991).

This is perhaps the epitome of an objectivist view of learning. The content of instruction is almost totally divorced from learning activity except as it is related to "processing" variables for "inputting" information. Learning is the input and mastery of particular content, the ability to repeat it, apply it, discriminate it where "if' is well defined. The ability of the person to think in the domain--to evaluate his understanding, judge relevance, and make decisions of what he needs toward what end--is irrelevant. Indeed, the irrelevance of thinking to the learning activity is reflected in this summary statement of the general view of learner control in instructional design and the state of our research findings:

The notion of learner control has long held intuitive appeal for developers of computer-assisted instruction, but its apparent potential for improving learning has never been experimentally established" (emphasis added, Goforth, 1994, p. 1).

And what is the intuitive appeal? It would seem that the belief is that learners should know best what they need, and so learning will be more efficient if they are in control. It has nothing to do with the thinking process being an integral part of knowing. As Ross and Morrison (1989) note, the notion of giving the learners control of their learning activities is based on two assumptions: learners know what is best for diem, and they are capable of acting appropriately on that knowledge. If the learner does not meet either of these assumptions, then control of "learning" is given to the computer so that learning can occur "efficiently."

The concept of learner control is similar to the concept of teacher-centered instruction in the noncomputer environment. That is, rather than supporting learners in developing control of their own learning and hence of being able to think in a domain, the teacher-centered and computer-controlled instructional approaches take responsibility away from the learner. However, in teacher control, it is primarily a control of the content and the basic learner task. In computer-control literature, the control is far more pervasive in that the computer takes over even the minute decision making. We find the title of a recent paper on control instructive as to the importance of this variable on the dynamics of a learning environment: I Lost Control (and My Students Found It) (Schleper, 1993).

7.4. 8 Assessment

Traditionally, assessment is an activity undertaken after learning is accomplished: Communicate some knowledge, then test to see if the knowledge has been successfully stored by the learner; demonstrate and coach a skill, then test to see if student can perform skill, etc. A great deal of technology of testing is devoted to enhancing the congruence of the testing context and the learning context. That is, is the test a reliable and valid measure of the extent to which learning has occurred? This approach also seeks to minimize factors that could contaminate test results. For example, a test that simply required the student to repeat answers to questions asked during learning would not be valid. Likewise, if the student were assisted in any way in completing the test, by a person or tool like a calculator, the results would likely be judged as invalid.

Generally speaking, the world of educational measurement adopts physical measurement as a model where a measurement tool quite different from the variable itself is applied: A ruler is applied to measure height, a micrometer is applied to measure thickness, etc. By analogy, an intelligence test measures intelligence but is not itself intelligence; an achievement tests measures a sample of a learned domain but is not itself that domain. Like micrometers and rulers, intelligence and achievement tests are tools (metrics) applied to the variables but somehow distinct from them.

The situation within constructivism is quite different. A rapidly growing literature (e.g., Shavelson, Baxter & Pine, 1992; Belak, Newman, Adams, Archbald, Burgess, Raven & Romberg, 1992; Gifford & O'Connor, 1992; Mabry & Stake, 1994; Linn, Baker & Dunbar, 1990) is introducing such terms as performance assessment, portfolios, authentic assessment, etc., and beginning the process of building a technology of assessment based on constructivist principles such as those proposed in this paper. 'Me distinction between learning and testing is certainly blurred if not rejected in these attempts. For instance, Ann Brown and her colleagues (1993) describe "dynamic assessment," a procedure that shares many characteristics with reciprocal teaching. It too is an example of "assisted learning" (see also Tharp & Gallimore, 1988), where the assessor establishes a zone of proximal development with the student to scaffold new learning as well as assess. Later in the assessment/teaching process, the teacher/assessor withdraws prompts when it is felt that the child can perform independently. Thus, if learning is in the connections, in the activity itself, then learning is the test. If the aim of a reading teacher, for example, is to have the child develop the skill of asking questions about reading materials, then the "test" is embedded in the teaching/ beaming context: Can the student now ask effective questions while reading, whereas previously she was only able to do so with the teacher's support and scaffolding? Or if the aim of the medical school faculty is to have students diagnose and prescribe treatment (and be able to defend their decisions), the test is embedded within the activity, not distinct from it. Ironically, perhaps, when traditional measurement techniques are used in situations such as these, where their applicability is questionable, the scores often show performance at least as good as traditional instruction (Albanese & Mitchell, 1993; Hubbard Welsh, latridis, Ficklin & Vaughn, 1994).

Additionally, performance assessment specialists are beginning to develop methods of large-scale assessment of complex performances (e.g., teacher certification--see Delandshire & Petrosky, 1994) that are authentic, discursive, semiotic, and reflexive. It is becoming increasingly clear that assessment that is sympathetic to constructivist principles will require new conceptions of such traditional concepts as reliability and validity (see, for example, Messick, 1989) and of the sorts of evidence that will be helpful in making assessment decisions like who should be employed, hired, admitted to university, and so forth. Of all the areas we have identified, this may be the one that is most underdeveloped (and under development) at this point in time.

7.4.9 Collaborative Learning

Collaborative learning and cooperative learning (see Chapter 35): Everyone wants it. It is the instructional strategy, perhaps the strategy of the decade. But why do we have students working together in groups? This is perhaps an area where one's metaphor for learning can most clearly be seen in its impact on the implementation of a strategy. The use of groups may simply be used as an alternative instructional strategy, with little change in the learning goals from traditional didactic instruction (Slavin, 1990; Johnson & Johnson, 1990). The guidelines for using cooperative groups focuses on structural and management variables like the gender distribution, number of participants, etc., and how to ensure that everyone does the work. From this perspective, groups are used for reasons that include providing variation in the classroom activity, teaching students how to cooperate and work together, sharing work loads and hence permitting larger projects, and to promote peer tutoring.

Groups also work on problems in the constructivist environment, and the goal in that work is to share alternative viewpoints and challenge as well as help develop each alternative points of view (Cunningham, Duffy & Knuth, 1993; Savery & Duffy, 1995; Sharan & Sharan, 1992). As we noted previously, learning is an inherently social-dialogical process. Hence, our reason for using groups is to promote the dialogical interchange and reflexivity. Our emphasis in providing guidance on the use of groups is how to promote that dialogical interchange among group members. We emphasize the importance of supporting collaborative informal reasoning about problems and reflectivity on the learning process.

7.4. 10 Computers and Media

Any Rip van Winkle who has just awakened after 10 years would undoubtedly be overwhelmed by the incredible changes in both the character and the pervasiveness of technology in our society. Video has moved out of the 11 professional production" limitation and out of the television and movie theaters to become a general medium available for the viewing and analysis of any event. The latest multimedia computers can store enormous amounts of information, present it via sound, text, video, graphics, etc., interact with users in modes that seem evermore natural and complex, and accomplish this with information and people distributed worldwide.

Given the widespread adoption of this technology in education and training, we will consider its role in education in some detail. Most often, technology is adopted by teachers and instructional designers as a "teaching tool," that is, to provide more effective and efficient delivery of instruction and hence more effective and efficient learning. According to this view, the video medium provides richer examples of concepts and principles, and thus we are able to teach the learner better how to execute a procedure, teach to discriminate between examples and nonexamples, etc. Computer technology permits us to build student and expert models (see 19.2.3) so that we can more effectively present problems to the learner and identify and remediate misconceptions (Psotka, Massey & Mutter, 1988). And, of course, the richness of the technology permits us to provider a richer and more exciting (entertaining-) learning environment that will better engage the student in learning the material being presented. We, on the other hand, want to focus on the technology as a tool for the learner rather than as a tool for the teacher.

Let us hasten to clarify, however, that we do not mean to simply substitute "learner" for "teacher" in the previous paragraph: The computer is not, or not only, a tool for the learner to acquire the content or skill more efficiently. Rather, our concern is the new understandings and the new capabilities that are possible through the use of technology. Pea (1985, 1993), in contrasting these two views, describes the first as using technology as a tool simply to amplify what we were doing before (so that we can do it more efficiently and effectively), while the latter is seen as augmenting cognitive activity and thereby leading to a reorganization and extension of our cognition.

One impact of the augmentation view is to examine how the tools may permit the learner to attend to higher-level representations by "off-loading" basic cognitive demands. For example, the use of the word processor permits easy reorganizing of text and hence permits the writer to explore alternative organizations, ways of expressing ideas, etc. It also dramatically impacts the nature of the interaction in collaborative writing activities. Similarly, the calculating functions permit many new foci in mathematics teaching (NCTM, 1989).

In addition to off-loading basic cognitive tasks, the technology may offer genuinely new representations or views of phenomena that would not otherwise be possible, and hence provide new understandings. Pea (1993) has noted that contribution of scientific visualization techniques to the understanding of particular phenomena (see, e.g., Keller & Keller, 1993). Hay (1994) and Soloway, Guzdial, and Hay (1994) are bringing those visualization techniques into the classroom as a strategy for aiding learners in developing new and richer representations of scientific as well as everyday phenomena. In a related vein, the Vanderbilt group has used video to capture complex activities in ways that allow learners to analyze those complexities and examine the interrelationships (CTGV, 1992). We might also point to the impact of video technology (the ease of recording as well as the ease of random access to and annotation of the video records) on our understanding of dynamic events like teaching and small-group collaboration (Jordan & Henderson, 1994; Brown, 1994).

Within this "augmentation" view, the MAB and MAR models of cognition offer different interpretations or understandings of the technology. The "mind as brain" view, with its focus on the individual mind, sees the computer as enhancing the individual's cognition, focusing on what he or she "knows." As Salomon, Perkins, and Globerson (1991) describe it, it is the effects of computing on the individual's cognitive skills that will impact cognitive performance outside of the computing environment. The effect of computing is an effect that endures beyond the computing, and, as such, this view suggests that knowledge (the residue effect) resides outside of the activity; it is an entity in the head.

From a MAR viewpoint, the technology is seen as an integral component of the cognitive activity. As discussed previously, cognition is distributed in the environment such that an understanding of cognition requires an examination of the activity in the environment. Bateson's famous example of the blind man provides what is to us a very clear example of distributed cognition: how it is impossible to separate tools from cognition:

Suppose I am a blind man, and I use a stick. I go tap, tap, tap. Where do I start? Is my mental system bounded at the handle of the stick? Is it bounded by my skin? Does it start halfway up the stick? Does it start at the tip of the stick? (1972, p. 459).

The answer to this, of course, depends on the activity of the blind man. In the case of navigating the world, the stick is an integral part of the cognition. However, when Bateson's blind man sits at a computer, the stick's relationship to his cognition has totally changed, and 'it is keyboards and mouses that become relevant. Cognition is distributed among the artifacts in the activity.

This view of distributed cognition significantly impacts how we think about the role of technology in education and training.8 The focus is not on the individual in isolation and what he or she knows, but on the activity in the environment (see Table 7-1). It is the activity-focused and contexualized-that is central. Furthermore the task of the learner is no longer seen as static-the computer as applied to the task-but rather it is dynamic: The computer opens new opportunities and makes available new learning activities. As Pea (1993) has noted, our goal becomes one of expanding cognition, not of reallocating cognitive activity as a division of labor.

A good example of the use of technology to expand cognition is found in George Landow's (1992) argument for the design and use of hypermedia (see 21.4). Landow has been perhaps the most outspoken advocate of hypermedia technology, for he sees it as a medium that permits critical theorists to realize and test their views as never before possible. His views on the important relation between hypertext and critical theory is clear in the following statements:

A paradigm shift, I suggest, has begun to take place in the writings of Jacques Derrida and Theodor Nelson, of Roland Barthes and Andries van Dam. I suspect that one name in each pair will be unknown to most of my readers. . . . [However] all four, like many others who write on hypertext or literary theory, argue that we must abandon conceptual systems founded on ideas of center, margins, hierarchy, and linearity, and replace them with ones of multilinearity, nodes, links, and networks. Almost all parties to this paradigm shift, which marks a revolution in human thought, see electronic writing as a direct response to the strengths and weaknesses of the printed book. This response has profound implications for literature, education, and books.... Using hypertext, we will have, or now already have, a new laboratory ... in which to test their ideas.... [While] critical theory promises to theorize hypertext, hypertext promises to embody and thereby test aspects of theory, particularly those concerning textually, narrative, and the roles or functions of reader and writer (1992, pp. 2-3).

... hypertext has much in common with some major points of contemporary literary theory and semiological theory, particularly with Derrida's emphasis on de-centering (see 10.5) and with Barthe's conception of the readerly VS. the writerly text. In fact hypertext creates an almost embarrassingly literal embodiment of both concepts (1992, pp. 33-34).

In essence, Landow is arguing that hypertext can be used to empower the reader to see and use text in new ways and in particular to support multilineal thinking. A goal of critical theory is to permit the reader his or her own center for investigation, not only in terms of starting points but also in terms of the consideration of the information and the paths along which those considerations might lead (see 9.2, 10.5.3). Thus a major goal in critical theory is to decenter the author and the text, to unconstrain the linkages of ideas from the linear flow of text and from the "container" of the book covers, and to place the authority for constructing and evaluating ideas in the reader and his or her collaboration with other readers. We would suggest, with the proliferation of information in this information age, that such decentering will become essential to successful problem solving and thought in many domains. Success will increasingly depend on exploring interrelationships in an information-rich environment rather than on accepting the point of view of one author who pursued one set of relationships and presents conclusions reflecting his or her implicit biases. While we tend to think of books as "natural" ways of representing information and ideas, Landow reminds us that it is an artificial structure that may not serve our present needs:

"The structure of books," Tom McArthur reminds us, "is anything but 'natural'-indeed, it is thoroughly unnatural and took all of 4,000 years to bring about. The achievement of the Scholastics, preeminently among the world's scribal elites, was to conventionalize the themes, plots, and shapes of books in a truly rigorous way" (Landow, 1992, p. 57).

In essence, the design of the text imposed order on fragmented knowledge and ideas. Hypertext would remove the textual imposition of order, and, Landow argues, the reader would create his or her own order based on scientific, historical, cultural, or any other thematically coherent focus. The consistency of Landow's view of hypertext with our constructivist MAR model is clear:

The hypertextual dissolution of centrality, which makes the medium such a potentially democratic one, also makes it a model of a society of conversations in which no one conversation, no one discipline or ideology, dominates or founds the others. It is thus the instantiation of what Richard Rorty terms the edifying philosophy, the point of which is to keep the conversation going rather than to find the objective truth" (Landow, 1992, p, 78).

Within Landow's framework, the diversity of "multilinearity" of a hypertext is critical, as is the availability of search tools and the ability of the reader to create his or her own links and nodes so as to find and create links and nodes in the process of constructing reader-centered themes. Landow (1989a, 1989b) illustrates the educational realization of his view of hypertext in the teaching of his undergraduate and graduate courses in literature where students work in a hypertext database consisting of thousands of nodes and linking to other hypertext databases. Student's assignments establish themes for which they must construct alternative interpretations, e.g., as to how two authors or two passages are similar, which may involve issues of the social or cultural characteristics of the time, early experiences of the authors, the formalities of plot development, etc. Rather than there being a "true" reason or interpretation, Landow encourages his students to recognize the multicausality and the importance of focus on interpretation.

Landow's use of hypertext in his literature courses provides an example of the use of technology to augment cognition and in the evolution of the learning task that the technology permitted. This is a considerably different view from that of a traditional "instructional" view of hypertext where the concern is whether the student will cover the material and where tools are designed to restrict access until prerequisites are "covered" or to guide the student through prespecified paths. In concluding this discussion, we would like to describe briefly two additional examples of this "constructivist" view of effective uses of technology.

First, a most obvious example is the use of the Internet and other wide-area networks to promote collaboration (see 14.1, 14.2, and 14.7). We have multiple examples of the use of the network to create international as well as scientific collaboration to help students take a less parochial view of issues, to help them to see and evaluate multiple perspectives, and to engage them in more authentic research activities (see, e.g., Roupp, 1993). On a more local level, Harasim (1993) has demonstrated the use of the Internet as a vehicle to promote collaboration among students in a distance education course. Most distance education programs use technology, if it is available, to deliver instruction, emphasizing the transmission of the content rather than collaboration. In contrast, Harasim made the dialogic central in her distance education course. Materials were distributed via the mail, and the Internet was used for formal seminar discussions and for informal exchanges in a virtual café. Just as would be expected in any on-site seminar, students were expected to discuss the seminar topic, and grading was based in part on their contribution to that discussion. There was no formal, didactic instruction.

Finally, Strategic Teaching Framework,9 (STF) provides an example of the use of multimedia in teacher education that honors the teacher as problem solver or researcher (Duffy, 1995). The goal of STF is to aid teachers in adopting a learner-centered, problem-solving, collaborative approach to learning, i.e., an approach based on constructivist principles. Typical technology-based approaches to teacher change use video to present alternative scenarios that the teacher can respond to and then receive the "correct" response, or scenarios illustrating "the" correct method by examples and nonexamples. STF, in contrast, views teaching holistically rather than as a set of discrete methods. Additionally, it is designed under the assumption that adopting a learner centered view requires a conceptual change in addition to the development of new strategies. Finally, it was assumed that teachers must construct and test their own understanding of methods and that construction must arise from their own evaluation of teaching.

This framework led to the design of STF based on the metaphor of visiting an ongoing classroom. If a teacher wanted to adopt new approaches to teaching, he or she would visit the classroom of experienced teachers, observe their teaching, ask questions, and explore different aspects of their approach. There would be multiple points of view expressed about what was important in the teaching process, and the teacher-learner would have to evaluate those perspectives. The teacher-learner would return to his or her classroom, test the strategies and views, and then, in an iterative process, return to the "mentors" classroom to observe more and ask additional questions. Ideally this teacher would be part of a community of teachers, all attempting to restructure their teaching, and the constructive dialogue would occur among members of this community.

STF, then, does not teach; rather it is a resource for learning. It consists of videos of classrooms the teacher-learner can visit. These are not brief scenarios, but rather an entire class episode, typically 50 minutes long. As teacher-learner sit in on the class, they may ask for points of view as to what is important instructionally at any particular point in the video. The learners have multiple perspectives available: the teacher they are observing, an experienced teacher-educator, or a researcher. Furthermore, the comments may address management, teaching, or assessment issues. Just as with the classroom video, none of the perspectives is scripted. They are straightforward comments from each "expert' as to what she or he thinks is important at that particular point in the class. Thus, this is an authentic interaction, both in visiting the classroom and in soliciting perspectives. While there is more to STF (a whole library of resources), this brief description illustrates the use of multimedia and, in particular, the richness of video, as a tool for the learner focused on constructing understanding rather than as a tool for the teacher (or instructional developer) to transmit knowledge.