Situated Cognition in Theoretical and Practical Context Brent G. Wilson and Karen Madsen Myers University of Colorado at Denver Final reference: Wilson, B. G., & Myers, K. M. (2000). Situated cognition in theoretical and practical context. In D. H. Jonassen & S. M. Land (Eds.), Theoretical foundations of learning environments (pp. 57-88). Mahwah NJ: Erlbaum. French translation: http://www.rezozero.net/articles/wilson_sitcog.htm The discipline of instructional design (ID) has always depended heavily on psychology for its theoretical grounding, often to the neglect of other important theory bases. Educational psychologists and cognitive scientists periodically turn their attention to instructional design, engaging the field in conversation and leading toward exciting new conceptions of learning and instruction. For some time, the brand of psychology being discussed in ID circles has been fairly progressive. Thus, while many of the systematic procedures for instructional development remain rooted in "efficiency" notions of curriculum, the newer instructional theories themselves tend to rest on constructivist and situated foundations. Consistent with this progressive shift, attention has been given to more open instructional metaphors such as learning environments and learning communities. The situativity movement in particular exhibits a rigor and broad disciplinary base that is promising as a foundation for thinking about learning environments (Young, 1993). How does situated cognition (occasionally termed "SitCog" in this chapter for convenience) differ from other learning theories that have influenced ID practice over the years? What new design implications does the theory hold for learning environments? More fundamentally, what role should any psychological theory play in the design of learning environments? These are the questions we address in this chapter. Our focus is on situated cognition, presented within a broader historical context that includes behaviorism and symbolic cognition, two other learning theories that have significantly shaped ID practice. At the outset, we should share some recommendations for designers and participants in learning environments: 1. We need to look beyond psychology-based learning theories and seek out perspectives from anthropology, critical theory, political science, etc. Because of its broader foundation, the situativity movement can be seen as an attempt to do just that. 2. In designing or participating within learning environments, we should take care not to apply any particular theory too dogmatically or uniformly. Toward the latter sections of the chapter, we outline a way of reading SitCog theories to encourage more ad hoc, situated use of theories toward the design and creation of learning environments. In writing the chapter, we faced a tension between conveying ideas in a technically adequate way and suggesting implications for practitioner audiences. To preserve a sense of the SitCog literature, we quote at length from the writing of various theorists. But while we value both goals of accuracy and utility, we feel a greater obligation to show how SitCog ideas might be useful as an organizing metaphor to guide the design and creation of learning environments. Thus some theoretical details are omitted, and other ideas are conveyed in language geared more for practitioners. SitCog at a Glance Situated cognition is a research approach, spanning many disciplines and objectives, that relates social, behavioral/psychological, and neural perspectives of knowledge and action (Clancey, 1997, p. 343). Lave (1991, p. 84) clarifies: ‘Situated’ does not imply that something is concrete and particular, or that it is not generalizable, or not imaginary. It implies that a given social practice is multiply interconnected with other aspects of ongoing social processes in activity systems at many levels of particularity and generality. Thus situated cognition should not be characterized as only allowing for concrete learning in localized situations. Instead, situated cognition emphasizes the web of social and activity systems within which authentic practice takes shape. Situated cognition comes into clearer focus when contrasted against traditional information-processing views of cognition. Don Norman (1993) does a good job conveying the essential difference between these two views. In Norman’s intentional caricature, traditional symbolic processing focuses on neural mechanisms and symbolic representations of mind: All the action is inside the head, yielding a natural distinction between the stuff out there and the processes taking place inside here. What could be more natural than to study the human by recognizing that the brain is the computational engine of thought, and thereby concentrating one’s efforts upon understanding brain mechanisms and mental representations? Seems pretty obvious. Sure, there is a lot of action in the world at large and within sociocultural groups, but cognitive processing occurs within the heads of individuals. So, all we have to do is understand the internal mental processes and the nature of the input/output transformations of individuals, and we will have covered everything that matters. (Norman, 1993, pp. 3-4) 2 In contrast, SitCog focuses on "the structures of the world and how they constrain and guide behavior": Human knowledge and interaction cannot be divorced from the world. To do so is to study a disembodied intelligence, one that is artificial, unreal, and uncharacteristic of actual behavior. What really matters is the situation and the parts that people play. One cannot look at just the situation, or just the environment, or just the person. To do so is to destroy the very phenomenon of interest. After all, it is the mutual accommodation of people and the environment that matters, so to focus upon only aspects in isolation is to destroy the interaction, to eliminate the role of the situation upon cognition and action. (p. 4) Table 1 offers an advance look at our views toward behaviorism, information-processing theory (or more broadly, symbolic cognition), and situated cognition. SitCog can be approached from a perspective that remains committed to understanding individual cognitive mechanisms. This is the approach taken by some artificial intelligence (AI) researchers and cognitive scientists, whose root interest lies in the nature of mind. Alternatively, SitCog may be seen from an almost entirely social or cultural vantage point. Jean Lave, for example, is loathe to acknowledge mental constructs and events. Knowledge is not an object and memory is not a location. Instead, knowing, learning, and cognition are social constructions, expressed in actions of people interacting within communities. Through these actions, cognition is enacted or unfolded or constructed; without the action, there is no knowing, no cognition. SitCog in this sense stands at the fringes of psychology much like behaviorism, with both approaches seeking to avoid mental constructs, focusing instead on the context or environment of actions and behaviors. Table 1. Two distinct traditions of situated cognition may be discerned: One with an individual focus on mind, the other with an external focus on community. Locus of Knowing What is Being Studied External or Group Internal or Individual Discrete events and Behaviorism Information-processing information (meaning theory, or cognition with a derived from the outside in) focus on symbolic computation Culturally and physically Situativity with a focus on Situativity with a focus on embedded activities and community participation and mind, neural, and physical meanings (meaning derived cultural construction, studied embodiment, studied by from the inside out) by anthropologists and cognitive scientists and AI ethnographers researchers In spite of these two discernible strands, situated cognition is positioned to bring the individual and the social together in a coherent theoretical perspective. This potential 3 integration of individual and social/cultural levels of scale is a primary strength which we discuss at length in the chapter (see also Chapter 2, Eds). Overview of Foundational Learning Theories We review in this section the main psychological foundations underlying the practice of instructional design, and by extension, the design of learning environments. Differences between these views can be understood in historical terms as part of evolving and competing paradigms for learning (Bruner, 1990; Cole, 1996; Mayer, 1992, 1996; Salomon & Perkins, 1998; Sfard, 1998). Behaviorism As Mayer (1992, 1996) has noted, behavioral psychologists such as Watson and Thorndike saw learning primarily as the acquisition and strengthening of responses. This view came to dominate American psychology throughout the first half of the 20th century. Because instructional design evolved out of educational psychology, this also was the view of most founders of the movement (e.g., Baker & Schutz, 1971; Gagné, 1965; Glaser, 1965; Mager, 1962; Markle, 1965). Nowadays, behaviorism is often dismissed as a serious theoretical stance for learning and instructional design, or given token attention. Even overview chapters such as ours apportion much less space to behavioral principles (cf. Mayer, 1992). Indeed, behaviorism is often shouldered with responsibility for status quo methods and conditions such as teacher-centered classrooms, lectures, and passive reception of material. This isn’t entirely fair. Recall that behaviorism was once a reform movement with a core commitment to active learning. To fully appreciate the contribution of behaviorism, we would need to understand, what were behaviorists trying to reform, and what did they bring to the table? Proponents of programmed instruction were dedicated to making instruction more individually tailored and effective in accomplishing its objectives. A full range of media and technologies were organized into new designs for instruction. Traditional methods such as teacher-centered classrooms and lectures were precisely what behaviorists were trying to reform. We also find it provocative to realize that the core assumptions of behaviorism bring the individual in close association with the environment. In fact, environment and behavior mutually define one another with action as the principal unit of analysis. The connection to situated cognition is evident. A full treatment of the influence of behaviorism on instruction and learning is beyond our scope (see Cook, 1997; Dick, 1987; Gagné, 1987; Hilgard, 1987; and Kliebard, 1987 for historical perspectives). For comparison purposes, and to suggest some concrete ways theory has indeed guided the design of learning environments, we have highlighted some key contributions of behaviorism to instructional design thinking (see Table 2). Although 4 these principles were developed before the current notions of learning environment gained prominence, they are useful for designing a particular form of learning environment that is common in many classrooms today (see Perkins, 1991 on different kinds of learning environments). Table 2. A sampling of behaviorist insights for the design of learning environments. Learn by doing. People learn best by actively engaging in tasks. This is commonly called "practice" or "learning by doing." Taxonomies. Learning outcomes can be differentiated in their type and complexity—for example, simple S-R bonds, concept classification, and rule-following. Such learning outcomes are compiled into classification schemes called learning taxonomies, which in turn guide selection of learning objectives and instructional strategies. Conditions of learning. For each type of learning, conditions can be identified that lead to effective learning. Identifying optimal conditions of learning forms the basis of prescriptive instructional theory, using the formula: To accomplish "X" learning outcome, apply or arrange for "Y" conditions. Behavioral objectives. Instruction should be based on clear, behaviorally specified learning objectives. Explicit formulation of objectives helps link instructional goals with evaluation and assessment, leading to increased accountability. Focus on results. Teachers and schools should be accountable for their students’ learning. Measurable behaviors are the best index of true learning outcomes and should be used to gauge instructional effectiveness. Alignment. Good instruction exhibits an alignment or consistency between learning objectives, instructional strategies, and strategies used to assess student learning. Misalignment of these components results in inadequate or unfair instruction. Task decomposition. People learn best when complex tasks are broken down into smaller, more manageable tasks and mastered separately. Prerequisites. Sub-tasks often become prerequisites to larger tasks. That is, students learn the larger task more easily when they have first mastered the sub-tasks. This leads to a parts-to-whole instructional sequence. Small successes. Sub-tasks have another advantage: They allow students to succeed. Succeeding at tasks is reinforcing, resulting in greater motivation to continue. Response-sensitive feedback. People learn best when they know the correctness of their efforts. When performance is not correct, specific information should be conveyed concerning what was wrong and how to improve the next time. Science of instruction. Educators need to be precise and systematic in their thinking, their 5 teaching, and their evaluation of students. Education can be treated as an applied science or technology, where through empirical inquiry, principles are discovered and applied. Performance support. People need support as they perform their jobs, through the use of job aids, help systems, and feedback and incentive systems. On-the-job, just-in-time training and support works best. In general, the closer the training is to job conditions, the more effective learning will be. Direct instruction. Giving clear directions, well-prepared presentations, suitable examples, and opportunities for practice and transfer—These are proven methods that result in substantial student learning. Pre testing, diagnostics, and placement. Students should not all be forced to endure the same instructional program. Instead, instruction should branch into alternative treatments according to prior skills, motivation, and other critical variables. Transfer. In order to be able to transfer a skill from one task to another, students need practicing doing it. If students never have opportunities to practice transferring their skills, they should not be expected to be able to perform on demand in test situations. We expect readers to respond differently to the enumerated principles. You may find some that you agree with, others that you feel strongly are outdated or damaging. Overall, their impact on education and training has been substantial, and many practitioners continue to rely on behavioral principles for designing learning environments in school and training settings. For reasons of economy and objectivity, behaviorism sought to explain learning in terms of observable behavior, generally avoiding reference to mental events and entities. This avoidance of mind and meaning was challenged, of course, but generally withstood challenges until the advent of cognitive science in the 1950s and 1960s. Information Processing Theory At the threshold of what would become known as cognitive science, the founders intended to bring mind to the center of psychology. Bruner (1990), a key participant, observes: [The] aim of [the cognitive revolution] was to discover and to describe formally the meanings that human beings created out of their encounters with the world, and then to propose hypotheses about what meaningmaking processes were implicated. It focused upon the symbolic activities that human beings employed in constructing and in making sense not only of the world, but of themselves. Its aim was to prompt psychology to join forces with its sister interpretive disciplines in the humanities and in the social sciences. (p. 2) 6 Concurrent with psychology’s emerging focus on mind, computer technology was developing. With the overlap of the computer and cognitive revolutions in both time and place, it was natural that of the potential driving metaphors, "one of the most compelling was that of computing" (Bruner, 1983, p. 274). Cognitive sciences ended up drifting away from the construction of meaning toward the processing of information, and the shift was metaphor driven (Bruner, 1990). Of course, there remained pockets of psychologists who continued to focus on mind and the construction of meaning; this less visible thread has always been there and can be traced back to Wundt’s work at the turn of the century. Information-processing theory (or symbolic cognition), while speaking a mechanistic language like that of behaviorism, held a number of advantages. Principally, the bottleneck caused by avoidance of mind was broken, resulting in a flood of new and productive research. Using methods such as reaction-time experiments, eye-movement studies, and think-aloud protocol, researchers were able to posit computational models of mind that filled many of the gaps left by behaviorism. We now had a dualistic framework: the world "out there" was represented by various memory structures inside the head. Our behavior within that outside world was thought to be accounted for by internal processes and mechanisms. Instructional designers could now think of learning in terms of taking experts’ cognitive structures, and mapping that knowledge into the heads of learners. The degree of similarity in cognitive structure between expert and novice was a good measure of whether learning objectives were being met. Table 3 presents key concepts from information-processing research that have a bearing on the design of learning environments. Table 3. Information-processing principles relating to the design of learning environments. Stages of information processing. Humans process information in stable, sequential stages, inputting sensory information into perceptual memory, then to working and long-term memory, and finally to response generation. In many ways, people are informationprocessing machines whose thinking and behavior can be modeled and simulated. Task modeling. Tasks can be modeled using flowcharts and other sequential representations. These models—called "cognitive task analysis"—can be used to pinpoint likely errors and make instruction more specifically targeted to the skill. Attention. Attention is often directed toward novelty or changes in one’s environment. Attention can suffer when instruction offers too much—or not enough—novelty, leading to anxiety or boredom respectively. Selective perception. Our goals, expectations, and current understandings color our perceptions. They serve as filters to the world, and shape our cognitive structures and responses. This selective nature of perception has implications for instructional sequencing, motivation, and metacognitive training. Memory load. We are only able to hold in mind about five to seven chunks of information 7 at a time. Problems arise when instruction taxes the limits of working memory. Major improvements can be made by careful analysis and revision of instruction to reduce memory demands. Memory-sensitive strategies include: sequencing instruction from simple to complex; allowing access to reference aids; and progressing in small steps with frequent repetition and elaboration. Kinds of knowledge. Two kinds of knowledge are fundamental: —Declarative knowledge (knowing that). Stored as propositions in semantic networks. —Procedural knowledge (knowing how). Stored as IF-THEN rules and pattern-recognition templates. These are often referred to in everyday language as knowledge and skill. Both declarative and procedural knowledge depend on representations of rules or information stored in memory. Skill compilation. Through repeated practice, skills become compiled or routinized. Several procedural steps are combined into a single whole, making performance easier and leaving cognitive resources available for other parts of a complex task. Talking about or un-learning a routinized procedure can be difficult, because details of task components are lost and have to be reconstructed. Automaticity is achieved when a second, simultaneous task can be performed without noticeable impairment of the first task. Meaningful encoding. Information is stored in long-term memory in ways that make it accessible for convenient retrieval. — Chunking. Information is chunked as it becomes organized into meaningful units, making it easier to remember. Chunked information fits together better and helps us overcome limits to working memory. —Elaboration. People make links between material and their prior knowledge through active thought and reflection. The more connections, the more meaningful and stable the item. Metacognition. Problem solving involves declarative and procedural knowledge, and something more. That we call metacognitive knowledge, involving self-monitoring, selfregulation, knowing when and where to deploy your strategies and knowledge. Motivation. Motivation is what makes people do what they do. Whereas behavioral explanations traditionally referred to instincts, drives, arousal, and reinforcement, cognitive theorists rely on models of cognitive processing and structure. Key concepts include incentives, self-efficacy, expectancy x value, success/failure attributions, performance versus learning goals, and intrinsic versus extrinsic motivation. 8 Experts versus novices. Experts differ from novices in a number of respects, including: —more domain-specific information to draw upon; —more refined domain-specific performance routines; —a commitment to steady periods of deliberate practice (reflective practice with the specific intent of skill improvement). Human development. Children’s growth in knowledge and skill can be interpreted as a series of stages from concrete to abstract forms of reasoning, or as accumulation of procedural and declarative knowledge about the world. Adults also grow in their epistemological understanding; this growth can also be characterized in terms of stages, moving from fixed, authoritarian views of knowledge toward views that acknowledge the important roles of interpretation and perspective. In both cases, instruction should be matched to development levels. Conceptual change. People make sense of their worlds by reference to schemas, mental models, and other complex memory structures. Differences between encountered experience and schemas can prompt further inquiry and reflection to resolve the conflict. Instruction should help learners assimilate and accommodate new information into existing schemas and cognitive structures. Like the previous table, Table 3 will strike readers in different ways. Most teachers and designers suffer an aversion to the computer metaphor. They don’t like to think of themselves as machines, or as belonging to the "same species" as symbol-processing machines. On the other hand, from our teaching experience, practitioners find in these concepts some concrete hooks for improving practice, e.g., managing memory load or prompting schema-based conceptual change. Over the years, it became increasingly clear that the information-processing metaphor provides a window into the mind, but a particular model of mind. According to this view, people (and computers) process information sequentially in a number of steps or stages. We selectively input information from our environment, then allow some of that information to be reflected on and acted upon: [W]e think before we act. [T]hinking goes on subconsciously, mediating every behavior. By this maneuver, we sought to distinguish human problem solving from uncontrolled, animal processes. (Clancey, 1993, p. 110) The meditation of thought is useful for understanding many cognitive tasks, but ultimately proves limiting as a complete framework. Do we really think about all our actions? How are we like other animals in our cognition? Are we following rules, or are the rules applied after the fact in explanations of our behavior? The reliance on rule-based models and explicit memory structures seems constraining and stridently dualistic. 9 Further, symbolic cognition focuses on individual processing, independent of cultural or physical context, as though "information" were a neutral construct. How do people’s cognition fit in with other people, tools, language, and culture? Surely tools, people, and culture are major mediators of constructed meaning (see also Chapter 5, Eds.). Situated Cognition Situated cognition has been positioned as an alternative to information-processing theory. It seeks to correct for some of the oversights of the symbolic-computation approach to cognition, in particular its reliance on stored descriptions of rules and information; its focus on conscious reasoning and thought; and its neglect of cultural and physical context. The terms "situated cognition," "situated action," or "situativity" enjoy no consensus among researchers. The terms are sometimes used to denote an array of related perspectives, something similar to "sociocultural constructivism." As we noted at the outset, at least two camps of researchers are typically associated with situated cognition: • Anthropologists like Jean Lave (1988, 1991) and Lucy Suchman (1993) are interested in the cultural construction of meaning. They meld anthropology and critical theory with the socioculturalism of Vygotsky (Forman, Minick, & Stone, 1993; Newman, Griffin, & Cole, 1989; Rogoff, 1990). The Vygotsky-inspired sociohistorical school takes as its central problem the processes whereby cultures reproduce themselves across generational boundaries. "This program provides a useful contrast to the behaviorist focus on low-level behavioral responses; and to Piagetian individualism on the other" (Kirshner & Whitson, 1997, p. 5). These researchers often avoid the term ‘situated cognition’ in favor of ‘situated action’ or ‘situated learning.’ (See also Chapter 8, Eds.). • Cognitive scientists like Allan Collins, John Seeley Brown, Don Norman, and Bill Clancey are interested in cognition at individual and social levels. Because these theorists tend to work from an AI or psychology tradition, individual cognition tends to draw more attention. For them, situated cognition has strong links to artificial intelligence, neuroscience, linguistics, and psychology, all fields with direct insights for understanding the individual mind. Other perspectives have been drawn to situated cognition, as illustrated by the wide variety of contributors to Kirshner and Whitson (1997). SitCog’s success in attracting researchers is another sign of its promise. Contributions from ecological and semiotic perspectives, in particular, are broadening and deepening the SitCog framework (see also Chapter 6, Eds.). Integrating Levels of Scale 10 The stand-out characteristic of situated cognition seems to be the placement of individual cognition within the larger physical and social context of interactions and culturally constructed tools and meanings. Situated cognition provides a "person plus" unit of analysis with individual knowing and social action intertwined. "The physical context is being reunited with the social, within the thought process" (Light & Butterworth, 1992, p. 1). Bredo (1994) depicts situated cognition as "shifting the focus from individual in environment to individual and environment" (p. 29). Suchman (1993, p. 72) makes this point by portraying Jean Lave’s research on cognition in everyday environments: [T]he very premise that schools constitute some neutral ground apart from the real world, in which things are learned that are later applied in the real world, is fundamentally misguided. [A]ll learning is learning in situ. [S]chools constitute a very specific situation for learning with their own cultural, historical, political, and economic interests: interests obscured by the premise that schools are asituational. Schools prepare students not for some generic form of transfer of things learned in schools to other settings, but to be students, to succeed or to fail, to more into job markets or not, and so forth. Of course, sociologists and critical theorists have been saying as much for years. The difference is that Suchman, Lave, and others are directly challenging the insulated view of cognition that ignores these contextual factors. Social and individual are not simply different levels of study—these levels interact and are inexorably connected. Bredo (1994, p. 29) describes this tight coupling: [Writing, conversing, or thinking are] the result of dialogue: in which person and environment (ideally) modify each other so as to create an integral performance. Seen in this way, a successful person acts with the environment, shaping it to modify himself or herself, in turn, and then to shape the environment, and so on, until some end is achieved. [T]he production of a well-coordinated performance involves a kind of dance between person and environment rather than the one-way action of one on the other. Designers of computer-based learning environments might argue that social and cultural factors play a minimal role, because the principal interaction is between computer and individual learner. Greeno (1997, pp. 9-10) disagrees: The situative view assumes that all instruction occurs in complex social environments. For example, a student studying along with a textbook or a computer tutor may not have other people in the same room at the time, but the student’s activity is certainly shaped by the social arrangements that produced the textbook or the computer program, led to the student’s being enrolled in the class where the text or program was assigned, and 11 provided the setting in which the student’s learning will make a difference in how the student participates in some social activity, such as a class discussion or a test. The principal challenge facing SitCog researchers is developing effective means of integrating levels of scale. The vision is clear, but the methodologies, both for research and for practice, are still in early stages of development. Until better tools become available, designers of learning environments should strive to be inclusive and expansive in their views, as they seek to integrate multiple perspectives in viewing the whole system rather than individual levels of their design thinking. Constructing Meaning As we mentioned above, situated cognition also challenges the orthodoxy concerning how we construct meaning. Researchers in AI and perception have generally taken a neural or connectionist approach to this question. Researchers like Winograd and Flores (1986) began taking seriously the criticisms of Hubert Dreyfus (1979). The result was an internal challenge to stage- and rule-based processing from within the AI community itself. Bill Clancey has been among the most vocal questioners of stored representations in memory, and rule-based knowledge systems. In the passage below, for example, Clancey argues that reflections, like other actions and thoughts, are immediate constructions: Every act of deliberation occurs as an immediate behavior. That is, every act of speaking, every motion of the pen, each gesture, turn of head, or any idea at all is produced by the cognitive architecture as a matter of course, as a new neurological coordination (Clancey, 1993, pp. 111-112). Clancey made some strong statements against stored representations and symbolic processing, for which he received some criticism. His most recent book (Clancey, 1997) takes a more moderate and consolidating approach, leaving room for conscious reflection and representations. St. Julien (1997, p. 7) illustrates how facts, rules, and representations might play a role in a situated framework: People who are competent in a field, people we usually think of as experts or simply as competent, are the producers or such facts. Facts, rules, and features are first used by experts among experts. They are helpful in discussing marginal cases, cases where the usual fluid, unremarkable competence has broken down. Facts, rules, and features provide the socially agreed on framework within which the fully competent can support each other when working on a difficult problem. While working within a situated framework, the roots of this position lie squarely within mainstream cognitive science research on concept learning. The difference is that the focus has shifted from individuals to groups and communities making use of shared rules and facts. 12 The debate about rules and plans has implications back to the role of instructional designer. Clancey (1995) contrasts two attitudes toward design: One rationalist or scientific, the other interactional. A rationalist view sees design in procedural or rulebased terms, with plans preceding action. You articulate the ideal up front, specify a design metric, and make adjustments for deviations as you proceed. Learning occurs through feedback regarding deviations and how adjustments are translated into redevelopment. Advocates of a systems design model for instructional development should see the parallel. Contrast that with the interactional view: What we do is created in our interactions; our representations orient action but don't control action. Representations include plans, formulas, algorithms, rules, and architectures. "How the plans are interpreted is itself an interactive, non-predictable process. Every behavior is an improvisation." (p. 27). "Human behavior is inherently ad hoc, inventive, and unique" (p. 28). "People do not simply plan and do. They continuously adjust and invent. Managing this process means managing learning, not managing application of a plan" (p. 39). Clancey's (1995) interactional view has specific implications for designers: • We must go into the community of the practitioner, using ethnographic methods of observation and reflection, and become participant observers. We develop a focus on how the community learns (pp. 33-34). • We must use methods of participatory design in which the worker participates in redesign practices with the designer (p. 38). The discussion on design illustrates the differences between symbolic-processing and situated cognition. The discussion also has further implications for how instructional designers approach the task of creating learning environments. A fuller discussion of these issues is presented in later sections of the chapter. Meaning Construction as a Social Activity Thinking and meaning construction are also approached from the social/cultural camp of SitCog theorists. On this view, construction of meaning is tied to specific contexts and purposes. We develop shared ways of responding to patterns and features in particular contexts that Gee (1997) calls discourses. "Discourses are sociohistorical coordinations of people, objects (props), ways of talking, acting, interacting, thinking, valuing, and (sometimes) writing and reading that allow for the display and recognition of socially significant identities" (pp. 255-256). Within discourses, people can exchange thoughts and explanations. Explanations or theories about practice tied to sociocultural groups are called cultural models (Gee, 1992; Shore, 1996). Cultural models are not just held by individual participants, but reside also in the practices in which the group engages, the tools they use, and the contextual setting. Resnick, Säljö, Pontecorvo, and Burge (1991) view thought and reasoning as "inherently (and throughout the life space) social activities in which talk and social interactions are not just a means by which people learn to think, but also how they engage in thinking. 13 They might say that discourse is cognition is discourse. One is unimaginable without the other" (p. 2; italics in original). Discourses and cultural models may not be physically present in the immediate context, but their impact on meaning construction is nonetheless critical. We achieve a sense of continuity over the life span as we interpret immediate situations in terms of past situations we see as having relevance to the here and now. Thus an immediate situation can only be by reference to the history of participants and associated groups: We interpret a text or a situation in part by connecting it to other texts and situations that our community or our individual history has made us see as relevant to the meaning of the present one. Our community, and each of us, creates networks of connections (and disconnections) among texts, situations and activities. These networks of connections that we make, and that are made in the self-organizing activity of the larger systems to which we belong, extend backwards in time as well [as] outwards into the socialmaterial world. (Lemke, 1997, p. 50) Table 4 continues our parallel exercise, summarizing key insights from situated cognition for the design of learning environments. We confess that this table was tougher to construct than the other two: The approach is newer and in some ways, more ambitious. Moreover, it has not been popularized and disseminated in the way the older traditions have been. Table 4. Situated cognition principles relating to learning environments. Learning in context: Thinking and learning make sense only within particular situations. All thinking, learning, and cognition is situated within particular contexts; there is no such thing as non-situated learning. Communities of practice: People act and construct meaning within communities of practice. These communities are powerful repositories and conveyors of meaning and serve to legitimate action. Communities construct and define appropriate discourse practices. Learning as active participation: Learning is seen in terms of belonging and participating in communities of practice. Learning is seen as a dialectical process of interaction with other people, tools, and the physical world. Cognition is tied to action - either direct physical action or deliberate reflection and internal action. To understand what is learned is to see how it is learned within the activity context. Knowledge in action: Knowledge is located in the actions of persons and groups. Knowledge evolves as we participate in and negotiate our way through new situations. The development of knowledge and competence, like the development of language, involves continued knowledge-using activity in authentic situations. Mediation of artifacts: Cognition depends on the use of a variety of artifacts and tools, chiefly language and culture. These tools and constructed environments constitute the mediums, forms, or worlds through which cognition takes place. Problem solving involves reasoning about purposes in relationship to the resources and tools which a situation affords. 14 Tools and artifacts as cultural repositories: Tools embody the history of a culture. They enable thought and intellectual processes and constrain or limit that thought. They also provide powerful means of transmitting culture. Rules, norms, and beliefs: Cognitive tools include forms of reasoning and argumentation that are accepted as normative in society. Using a tool in a certain manner implies adoption of a cultural belief system about how the tool is to be used. History: Situations make sense within an historical context, including the past experiences and interactions of participants, as well as anticipated needs and events. Cultures, through tools, artifacts, and discourse practices, embody the accumulated meanings of the past. Levels of scale: Cognition can best be understood as a dynamic interplay between individual and social levels. Focus on one level, while assuming constancy or predictability at the other, is bound to at least partly misinterpret the situation. Interactionism: Just as situations shape individual cognition, individual thinking and action shape the situation. This reciprocal influence constitutes an alternative conception of systemic causality to the more commonly assumed linear object causality. Identities and constructions of self: People’s notion of self—of continuing identity, separate from others yet belonging to various groups — is a constructed artifact with many uses. People have multiple identities, which can serve as tools for thinking and acting. These principles are also less specific in their prescriptions for the design of learning environments. Even in their descriptive nature, however, the principles differ markedly from those in the earlier tables. A SitCog approach to learning environments would pay close attention to language, activities of individuals and groups, cultural meanings and differences, tools (including computer tools and environments), and the interaction of all of these together. Assessing the potential of a given learning environment would involve understanding how these components combine in a way that results in participation in activities valued by individuals and groups. How are we to evaluate the potential of situated cognition as an aid to the design of learning environments? Sfard (1998, p. 8) notes that we usually think of learning as "acquiring" something—perhaps a competency, skill, or capability. SitCog theorists have operated from a different metaphor, that of participating in communities of authentic practice. Lave (1988) argues that people are principally defined by their participation in an activity as well as the roles they assume in social practice (see also Lemke, 1997, p. 45). Sfard sees an upside to this participation metaphor: The promise of the [participation metaphor] seems, indeed, quite substantial. The vocabulary of participation brings the message of togetherness, solidarity, and collaboration. [It] does not allow for talk about permanence of either human possessions or human traits. Being "in action" means being in a constant flux. The awareness of the change that never stops means refraining from a permanent labeling. Actions can be clever or unsuccessful, but these adjectives do not apply to the actors. For the learner, all options are always open, even if he or she carries a history of failure. Thus, quite unlike the [acquisition metaphor], the [participation 15 metaphor] seems to bring a message of an everlasting hope: Today you act one way; tomorrow you may act differently. (Sfard, 1998, p. 8) This perspective affords greater flexibility in defining selves, roles and identities. In a learning environment, this flexibility could likely contribute to more positive risk-taking, inquiry, and appreciation of multiple perspectives. On the other hand, situated cognition has been criticized on a number of points, especially areas where theorists have overstated their positions (Anderson, Reder, & Simon, 1996). Sfard (1998), while sympathetic to situated approaches, acknowledges the difficulty situated theorists have in explaining transfer of knowledge from one setting to another: A persistent follower of the [participation metaphor] must realize, sooner or later, that from a purely analytical point of view, the metaphorical message of the notion of transfer does not fit into [their] conceptual frameworks. Learning transfer means carrying knowledge across contextual boundaries; therefore, when one refuses to view knowledge as a stand-alone entity and rejects the idea of context as a clearly delineated "area," there is simply nothing to be carried over, and there are no definite boundaries to be crossed. (Sfard, 1998, p. 9) Adult educators in particular face pressure to deliver training that can transfer to job situations. Situated theorists don’t speak their language if they can’t promise some transfer to the job. Tripp (1993, p. 72) highlights another potential downside of situated approaches, issuing a warning against leaving learners to their own resources: In general, studies of adults who have learned languages "on-the-job" reveal a phenomenon called "fossilization." Fossilization refers to the learning of incorrect, but understandable, syntax and pronunciation which suffices for communication. Since this interlanguage allows satisfactory social interactions, the learner does not progress to a higher degree of mastery and, thus, the mistakes are fossilized and become part of the learner’s permanent repertoire. To the degree that authentic environments create unsupervised performance opportunities, such learning variations are liable to happen. Some of these learning variations will be positive, while others admittedly negative. Designers of learning environments will want to carefully weigh the risks and benefits of authentic experiences, and provide supervision and guidance as appropriate. Communities can definitely convey wrong values, and individuals can fit in poorly or pick up messages that work against academic learning goals. Salomon and Perkins (1998, p. 21) note how interactions between individual and collective learning can go bad: 16 What is learned by an individual may upset or even subvert rather than abet collective ends, as with the student taking advantage of his or her team members’ work or the corporate climber being more interested in personal advancement than in the overall success of the organization. In such cases, the collective has "learned," but what it has learned happens to be profoundly limiting both for itself and for the participating individuals. Granting consideration to these various problems, we still see situated cognition holding great promise as a guide to the design of learning environments. In the next section, we explore how SitCog might successfully serve as an integrating framework for including multiple learning theories and levels of scale. Then we turn to reflections on the appropriate role of theory in the design of learning environments. Situativity as an Integrating Framework Norman (1993) posed a nicely competitive question: Does symbolic cognition accommodate situativity, or does situativity accommodate symbolic cognition? Greeno and Moore (1993) believe that a situated approach can best serve as an integrating framework: [W]e see, in the present situation, a prospect of completing a dialectical cycle, in which stimulus-response theory was a thesis, symbolic information-processing theory was its antithesis, and situativity theory will be their synthesis. In the 1950s and 1960s, when the theory of symbolic information processing was being developed, the prevailing stimulusresponse theory in psychology lacked resources for analyzing and representing the complex structures involved in mental activity. A goal of stimulus-response psychology was to account for behavior as much as possible in terms of externally identifiable factors, and the structures of information and procedures were contained in a theoretical "black box." The theory of symbolic information processing has allowed us to investigate the contents of that black box in detail. We contend that symbolic processing theory presents another black box that contains the structure of interactive relations between cognitive agents and the physical systems and other people that they interact with. (Greeno & Moore, 1993, p. 57) Part of what makes situated cognition a contender for integration is its ability, as we have discussed, to accommodate both individual and social scales of study. Salomon and Perkins (1998) recall the analogy of a spreading flu. Consider two levels of explanation— cell biology and epidemiology: Clearly, the two complement each other. Subverted cellular mechanisms figure in the invasion of individual cells by viruses, but the viruses have to arrive at individual cells to infect them. Although each process can be 17 understood in its own right, understanding the interplay yields a richer and conceptually more satisfying picture. (Salomon & Perkins, 1998, p. 2) The analogy sounds very appealing, but how far can competing theories, emphasizing different levels of scale, really fit together? How commensurable are models of individual cognitive processing with broader depictions of situated action? Greeno believes in the commensurability of views, arguing for accommodation of behavioral and information-processing strategies within an overall situated framework. This is possible, he says, because SitCog acknowledges the various contexts where such strategies may be needed and relevant: [L]earning environments organized on behaviorist skill-acquisition principles encourage students to become adept at practices, involving receptive learning and drill, that result in efficient performance on tests, and learning environments organized on cognitive knowledge-structure principles encourage students to become adept at constructing understanding on the basis of general ideas and relations between concepts. (Greeno, et al., 1998, p. 14) In other words, different situations will call for different tools, models, methods, etc. Instructional methods are thus seen as tools to be appropriated by participants within the local situation, rather than general prescriptions to be used in all learning situations. While learning goals and purposes should always be examined critically, the specific choices of goals and activities are rightfully placed within the specific situation. Greeno and colleagues contrast the three foundational learning theories, while maintaining the primacy of a SitCog framework: Behaviorist principles tend to characterize learning in terms of acquisition of skill. Cognitive principles tend to characterize learning in terms of growth of conceptual understanding and general strategies of thinking and understanding. Situative principles tend to characterize learning in terms of more effective participation in practices of inquiry and discourse that include constructing meanings of concepts and uses of skills. We argue here that the situative perspective, focused on practices, can subsume the cognitive and behaviorist perspectives by including both conceptual understanding and skill acquisition as valuable aspects of students’ participation and their identities as learners and knowers. (Greeno, et al., 1998, 14) Greeno is optimistic that elements of behaviorism, cognitivism, and situativity can be combined: Both the behaviorist skill-oriented and cognitive understanding-oriented perspectives have informed the development of educational practices 18 significantly, but they are often portrayed, in research literature and the popular press, as diametrical opposites, where learning according to one view precludes learning according to the other. We argue here that important strengths and values of behaviorist and cognitive practices can be included in practices on the basis of the situative principles of valuing students’ learning to participate in inquiry and sense-making. Situative principles can provide a useful framework for evaluating the contributions of behaviorist and cognitive practices in a larger context. (Greeno, et al., 1998, p. 15). Continuing: [I]n the situative perspective, both learning to participate in the discourse of conceptual meanings and learning basic routines of symbol manipulation can both be seen as significant assets for student participation, rather than being orthogonal objectives. (Greeno, et al., 1998, p. 17) We cautiously agree with Greeno. Situated cognition, because of its holistic tendencies and preferences for rich, active environments, is well-positioned to serve as a synthesis or integrating framework. It seems easier to imagine an open, holistic framework accommodating a more technical, symbolic framework than the reverse. Ultimately, we believe some way must be found to accommodate multiple levels of scale and, to some extent, competing paradigms or theories. Gould makes a similar point about the need to include both biology and culture: We must go beyond reductionism to a holistic recognition that biology and culture interpenetrate in an inextricable manner. One is not given, the other built upon it. Individuals are not real and primary, with collectivities (including societies and cultures) merely constructed from their accumulated properties. Cultures make individuals too; neither comes first, neither is more basic. (Gould, 1987, p. 153; cited in Clancey, 1997, p. 244). At this point, accommodating multiple levels of scale within a situated framework is not fully realized. Some camps demonstrate antipathy to a neural or psychological level of scale. We see no inherent incommensurability between a SitCog framework and many neural and information-processing concepts of individual cognition. This is an area where we would like to see further development and discussion; we are optimistic that a greater level of integration can take place between theories that are seen, at present, as competitors. Implications for Learning Environments Greeno offers what might constitute a mission statement for situated cognition as an integrating framework: 19 We need to organize learning environments and activities that include opportunities for acquiring basic skills, knowledge, and conceptual understanding, not as isolated dimensions of intellectual activity, but as contributions to students’ development of strong identities as individual learners and as more effective participants in the meaningful social practices of their learning communities in school and elsewhere in their lives. (Greeno, et al., 1998, p. 17) We share Greeno’s enthusiasm, yet a tension is emerging. Situativity can be read in two competing ways: • As a prescriptive basis for design; or • As a descriptive basis for understanding and accommodating different purposes, tools, methods, etc. On the one hand, we could look to SitCog theory to tell us what kind of methods to use to achieve active, thriving communities of practice. This view would champion certain methods as promoting learning (e.g., authentic performance in information-rich, authentic settings) while eschewing others (e.g., direct instruction or efforts to manage cognitive load). Barab and Duffy (this volume) present a good example of using the theory in this way. Alternatively, we could look to SitCog to help us understand, through observation and critique, how a given learning environment combines elements to accomplish certain goals. Is the learning environment successful in accomplishing its learning goals? How do the various participants, tools, and objects interact together? What meanings are constructed? How do the interactions and meanings help or hinder desired learning? At first glance, the second use of the theory may seem unrelated to design. We believe, though, that a descriptive approach, while not offering a direction path to action, holds some advantages for thinking about design. First, as Greeno has shown, a descriptive use of SitCog can more easily accommodate a variety of conceptions and purposes for instruction. Because such a view does not attach itself to a clear set of instructional methods, it will likely be relevant to more learning situations. Second, because descriptive critiques start with the given situation, designers and participants may more clearly see a path toward improvement. This may help to mitigate the concern expressed by some practitioners that ideologically based approaches are just too extreme or idealistic to adopt wholesale. Finally, a less prescriptive approach will likely be more respectful of participants’ values and concerns. Communities of practice represent diverse interests and political beliefs. Any learning environment meant to reflect that diversity would need to be able to incorporate varied methods, approaches, strategies, etc. The two contrasting uses of SitCog theory result in an interesting irony: 20 • The prescriptive use of SitCog, while pursuing authentic communities of practice, runs the risk of becoming just another model to be imposed on practitioners out of context and without regarding to situational concerns. At the same time: • The descriptive use of SitCog runs the risk of legitimating practices that run counter to authentic, situated learning. Rather than resolve this difficulty, we would hope to see some attention to both concerns. Care should be taken to respect local purposes and culture, while at the same time promoting positive values such as authentic activity within communities of practice. A further tension exists concerning how feasible it is to control and thus design authentic learning environments. If we are reading the situated people correctly, authentic communities of practice are not so much designed, but rather emerge within existing environments and constraints. They fill ecological niches where certain opportunities open up, based upon the environment, people, tools, organizational structure and power dynamics, etc. How can we talk about designing authentic learning environments when so much of what goes into a learning environment is predetermined by constraints, or emerges based on the participants themselves? Brown and Duguid present the issue another way. They cite a wonderful observation on the ineffability of knowledge: A very great musician came and stayed in [our] house. He made one big mistake: [he] determined to teach me music, and consequently no learning took place. Nevertheless, I did casually pick up from him a certain amount of stolen knowledge. (Rabdrath Tagore, quoted in Bandyopadhyay, 1989, p. 45; cited in Brown & Duguid, 1993, p. 10) Not only does this statement suggest the tacit nature of important knowledge; it also suggests the near impossibility of conveying knowledge systematically, e.g., by design. Nonetheless, using situated principles, Brown and Duguid have hope that learning environments can be designed that will succeed in conveying tacit but important knowledge: It is a fundamental challenge for design—for both the school and the workplace—to redesign the learning environment so that newcomers can legitimately and peripherally participate in authentic social practice in rich and productive ways—to, in short, make it possible for learners to "steal" the knowledge they need. (Brown & Duguid, 1993, p. 11) Again: The best way to support learning is from the demand side rather than the supply side. That is, rather than deciding ahead of time what a learner 21 needs to know and making this explicitly available to the exclusion of everything else, designers and instructors need to make available as much as possible of the whole rich web of practice—explicit and implicit, allowing the learner to call upon aspects of practice, latent in the periphery, as they are needed. (Brown & Duguid, 1993, p. 13). Brown and Duguid seem not so eager to accommodate behaviorist and cognitivist perspectives. Yet their vision of designing situated-learning environments is also compelling. There seems to be room within the situated camp for a variety of approaches to instruction and the design of learning environments. The problem of how to design something that seems un-designable is not resolved; however, the process of design seems one of coordinating and compiling resources, then guiding participation—as opposed to pre-specifying complete learning resources and activities. Further Questions About Theory and Practice In this section, we turn more directly to general questions concerning how theory and practice might best work together in the design and development of learning environments. Grounded Design Mike Hannafin and colleagues (Hannafin, 1997; Hannafin, Hannafin, Land, & Oliver, 1997) are fully aware of the competing paradigms for instructional design. Moreover, they see potential strengths in different theories of learning and instruction. As a response to the plurality of perspectives, they have developed the concept of grounded-learning systems design. Grounded-learning systems design is defined as "the systematic implementation or processes and procedures that are rooted in established theory and research in human learning" (Hannafin et al., 1997, p. 102; see also Chapter 1, Eds.). Following this model, ID practices (or approaches) should be grounded in some theory validated by some research tradition. Good behaviorally based instruction might coherently and consistently follow a Gagné-style approach, for example, while good situative instruction would be grounded on very different principles and values. Consistency is the key: A learning environment described as reflecting cognitive-information processing views of learning, yet failing to account for limitations in shortterm memory, reflects a mismatch between presumed foundations and assumptions and their associated methods. A constructivist’s learning environment that decontextualizes and tutors to mastery is equally ungrounded. (Hannafin, 1997) The test for legitimacy has thus shifted from using the right theory to grounding practice in the right way. Grounded practice for instructional design must be: 22 • Based on a defensible theoretical framework; • Consistent with research validating that theoretical framework; • Generalizable to other cases and situations; • Empirically validated through successive tryout and revision (Hannafin, 1997; Hannafin, et al., 1997; see Chapter 1, this volume, Eds.). The concept of grounded design is a clear attempt to accommodate multiple theories for designing sound instruction. Many designers may find a grounded approach more realistic than the search for the one true theory. The decision about an appropriate theory to use is thus situated in the local environment. We are somewhat sympathetic to this concept, especially its effort to negotiate a truce among theoretical positions. But we need some clarifying of the construct. The examples given by Hannafin et al., (1997) seem to be of two different sorts: • Schools and instructional programs in operation at some site (e.g., airplane pilot schools); and • Nodels and related materials designed by theorists to be more broadly used by practitioners (e.g., the Jasper series; the Knowledge Integration Environment or KIE; Microworlds Project Builder). These two types of product are somewhat different. The authors present an example of two private airplane pilot schools: one employing a directed-instruction approach and one more hands-on and situational (Hannafin et al., 1997, p. 107). These pilot schools are clear examples of instruction in real contexts. The theorist-designed models, however, are not so clearly grounded in specific sites or communities. While the models incorporate substantial elements of design and decision-making, we see them principally as tools to facilitate instruction rather than instruction itself. In the usual sense of the term, instruction happens when these conceptual tools and resources get used in real settings by real people. The instructional models cited by Hannafin are tools for the theorist (to test out and disseminate their theories), and tools for the practitioner (to appropriate and use them in specific settings). Returning to Hannafin’s example of the two pilot-training schools, we are left to wonder: Did the schools design their approaches by reference to particular theories, or did they arrive at their approaches without recourse to theory? We could imagine either scenario: Maybe the schools hired a consultant or a trainer schooled in a particular theory. Or maybe a veteran pilot saved her money, started a school, and developed the curriculum in the way that seemed best. Or more likely, the training approach evolved over time, undergoing occasional revision based on a number of influences, some theory-based, some intuitive, some reflecting constraints of the business or of the teaching situation. 23 The idea of grounded design is a helpful construct that relates specific instructional models back to more general theories of learning. The cognitive apprenticeship model should be consistent with its grounding theory of situated cognition. The Jasper Series and its associated model of anchored instruction should be consistently aligned with principles of authentic, constructivist learning advocated by its designers. Gagné’s nine events of instruction should be grounded in the behavioral/cognitive learning theory of its author. These specific models and materials provide a concreteness and directness that practitioners find extremely helpful in converting theories to practical action. On the other hand, we wouldn’t expect—or want—the same level of consistency between the details of real-life instruction and a given theory. In fact, we believe that most instructional programs in operation would be ill-served by a strong linkage to a particular theory. To explain our concerns, we return to a more general discussion of educational theories and their relation to practice. Educational Theory-making in Cultural Context Labaree (1998) contrasts educational knowledge with that gained in other sciences and professions. Educational research, he says, generates: • Soft (as opposed to hard) knowledge that is not easily verifiable, definitive, and cumulative; and • Applied (as opposed to pure) knowledge that is more local, less generalizable, and less theoretical. To clarify this point, Labaree offers an interesting analogy. Imagine knowledge in geographical/architectural terms. An "urban" discipline would accumulate knowledge in concentrated, central locations. High-rise buildings could be built on the strong foundation of secure, replicable findings. Hard sciences are like that. In contract, education more closely resembles a rural landscape, with farmhouses and outbuildings distributed across wide open spaces. Educational knowledge is hard to build up into strong, stable structures; rather, it seems very locally distributed across settings, problems, participants, and content areas. Shulman (1987) makes a similar point about education’s special epistemological status: Education is not a "science" in the sense of those sciences discussed by Popper, Kuhn, Lakotos, Feyerabend: [its] major focus is (or ought to be) on an artifact called "practice". It is the marriage of theoretical knowledge with practical action which characterizes education (along with medicine, law, and other "professional fields") and requires a philosophical perspective of its own (pp. 39-40). The soft, applied, "rural" nature of educational knowledge results in some predictable effects, some good, some bad: lower status within the academy; more flexibility in determining research problems; more recurring waves of reform; more diversity in 24 perspective; greater influence of outside theories and frameworks. One would hope, as Shulman suggests, for greater appreciation of practical knowledge. Too often, though, the knowledge and wisdom gained from years of practitioner experience is subordinated to the structured, formal knowledge of the university researcher or textbook. Practitioners are asked to "apply" the knowledge or theory developed by the researcher; the practitioner role is at risk of being reduced to that of technician. Theories in Situated Context According to Blum (1970, p. 303), "Aristotle used theoria as that kind of mental activity in which we engage for its own sake, as contemplative: to theorize meant to inspect or to keep one’s gaze fixed on. To theorize was to turn one’s mind in a certain direction, or to look at the world under the auspices of a certain interest." But since Aristotle, ‘theory’ has taken on a whole host of meanings. Thomas (1997) claims that the lack of precision in meaning has resulted in its being equitable to nothing more than "intellectual endeavor" (p. 75). Thomas’s (1997) critique speaks of the hegemony of theory and argues for more "ad hocery" and anarchy in thought and method: [T]heory of any kind is a force for conservatism, for stabilizing the status quo through the circumscription of thought within a hermetic set of rules, procedures, and methods. Seen in this way, theory—far from being emancipatory—is in fact an instrument for reinforcing an existing set of practices and methods in education. (Thomas, 1997, p. 76) Why is theory harmful? The answer is that theory structures and thus constrains thought. Thought actually moves forward, Feyerabend says, by "a maze of interactions: [by] accidents and conjunctures and curious juxtapositions of events (Feyerabend, 1993, p. 9). The naïve and simpleminded rules that methodologists use cannot hope to provide the progress for which we wish. He quotes Einstein as saying that the creative scientist must seem to the systematic epistemologist to be an "unscrupulous opportunist." Holton (1995) also draws on Einstein, saying that the essence of scientific method is in the seeking "in whatever manner is suitable, a simplified and lucid image of the world. There is no logical path, but only intuition (p. 168). In other words, Feyerabend (1993) concludes, "the only principle that does not inhibit progress is anything goes" (p. 14). (Thomas, 1997, p. 85) While we acknowledge a more positive role for theory use than Thomas does, we believe that creators and users of theory should be more aware of ways theory can be misused in practical situations. Theory-based or Theory-informed? 25 How then should practitioners make use of theory as a guide to the design of learning environments? A theory-centered approach runs the risk of putting theory in charge, with the practitioner subordinate to the ideas. This risk is most evident when we speak of "applying" theories to practice, as though theories were simple technologies to be applied. Bednar, Cunningham, Duffy, and Perry (1995) state: [E]ffective instructional design is possible only if the developer has reflexive awareness of the theoretical basis underlying the design: [it] emerges from the deliberate application of some particular theory of learning (pp. 101-102, cited in Hannafin et al., 1997). While these authors do not advocate a simple application of theory to practice, theirs is nonetheless a fairly cognitive/rational position. Instructional designers deliberately apply some theory of learning. Whatever decisions or strategies they choose should be consistent with this underlying theory. The Hannafin idea of grounded practice rests on this same premise. How else can we imagine the stance toward theory of designers and participants of learning environments? Most clinical psychologists are reportedly "eclectic" in their stance towards the various theories of psychotherapy. Many teachers and instructional designers take the same non-commital stance toward theory. They prefer a menu or toolbox metaphor instead of an application/consistency metaphor. Practitioners tend to be opportunistic with respect to different theoretical conceptions; they might try viewing a problem from one theoretical perspective, then another, and compare results. This stance toward theory might be termed "eclectic" or "grab-bag," but we prefer to think of it as problem- or practitioner-centered. People, rather than ideologies, are in control. The needs of the situation rise above the dictates of rules, models, or even standard values. As we have indicated, we believe such a person-centered approach may be the most situated of approaches, even when the resulting strategies do not look wholly "situated" or "authentic." Sfard (1998) highlights the dangers of narrowly applying a single theory to practice: When a theory is translated into an instructional prescription, exclusivity becomes the worst enemy of success. Educational practices have an overpowering propensity for extreme, one-for-all practical recipes. A trendy mixture of constructivist, social-interactionist, and situationist approaches is often translated into a total banishment of "teaching by telling," an imperative to make "cooperative learning" mandatory to all, and a complete delegitimization of instruction that is not "problem-based" or not situated in a real-life context. But this means putting too much of a good thing into one pot. Because no two students have the same needs and no two teachers arrive at their best performance in the same way, theoretical exclusivity and didactic single- 26 mindedness can be trusted to make even the best of educational ideas fail. (Sfard, 1998, pp. 10-11) Sfard seems to be arguing for precisely the opposite of tidy consistency. She continues: What is true about educational practice also holds for theories of learning. Dictatorship or a single metaphor, like a dictatorship of a single ideology, may lead to theories that serve the interests of certain groups to the disadvantage of others. When two metaphors compete for attention and incessantly screen each other for possible weaknesses, there is a much better chance for producing. a liberating and consolidating effect on those who learn and those who teach. (Sfard, p. 11) Good design of learning environments should be informed by theory, but not slave to it. Designers and participants may keep a theory in mind—or maybe many theories at once—when considering a problem and deciding on a course of action. But the problem is at the center, not the theory. Local conditions will recommend a solution—in the contextualized, contingent reasoning of professional practice, which cannot be captured by the technical rationality of abstract theories, research, and generalities. Respecting Real Communities Situated cognition treats culture as a powerful mediator of learning and practices, both for students and teachers. Think about a typical problem in practice: to a large extent, the surrounding culture and scales of community define what is possible and what is real. A school district applying a uniformly consistent a SitCog-inspired teaching method will run into trouble if the broader community’s values are not represented. Consistent theoretical grounding is only possible or desirable where participants share a common ideology. Examples might include the military, a small company, or a charter school. But even in these cases, constituencies have this maddening tendency to diverge off the beaten path; to seek innovation and change; to differ on even fundamental points. Resulting instructional designs are likely to be some sort of compromise, reflecting the diversity of participants and stakeholders. And rather than being seen as a weakness for lack of theoretical consistency, we tend to see such compromises as valuable reality checks. The hegemony of theory is resisted, and the needs of real people are accounted for. Through democratic and dialogical processes, a local solution is found to problems, synthesizing diverse interests, beliefs, and needs, hopefully crossing ideological boundaries to include the full community. Our critique of theory can be applied to instructional methods as well. We believe that instructional quality cannot be guaranteed by adherence to a particular method or strategy. Quality or effectiveness has as much to do with relationships, contexts, and situations, as it does with method. Quality has as much to do with how a method is realized, than with how the method is typed or categorized. Parker Palmer (1997, p. 16) put it this way: "Good teaching cannot be reduced to technique; good teaching comes from the identity and integrity of the teacher." We realize this challenges a core tenet of 27 traditional ID doctrine. Unfortunately, a full discussion of the limitations of method beyond the scope of this chapter. Summary and Conclusion We have laid out a number of claims in this chapter concerning situated cognition. First we presented situated cognition as the latest learning theory with potential for grounding learning-environment design, following behaviorism and information-processing theory. However, while SitCog holds some advantages over previous foundations, it does not presently offer a comprehensive account of cognition. "In its more specific implications for education, situated cognition theory has yet to refine a distinct and distinctive approach" (Kirshner & Whitson, 1998, p. 27). For SitCog to fully serve as an integrating framework, a means of accommodating multiple perspectives needs to be developed, to allow inclusion of selected ideas and practices from behaviorism, symbolic cognition, and other theories, both psychological and non-psychological. SitCog also presents an opportunity to define the designer’s role in new ways. The design task is seen in interactional, rather than rational-planning, terms. But more importantly, design and control become situated within the political and social context of actual learning environments. Rather than applying the best learning theory, designers and participants of learning environments honor the constraints and affordances of the local situation. The use of theory within such learning contexts becomes much less linear and direct. Like any tool, practitioners can find value in various theories, especially in providing alternative lenses for seeing problems. Theories, like other tools, help define the situation, and are in turn defined by them. A situated view of design, then, is one that supports the worthy practices of participants and stakeholders, using whatever theories, tools, or technologies at their disposal. References Anderson, J. R., Reder, L. M., & Simon, H. A. (1996, May). Situated learning and education. Educational Researcher, 5–11. Baker, R. L., & Schutz, R. E. (1971). Instructional product development. New York: Van Nostrand Reinhold. Blum, A. F. (1970) Theorizing. (pp.301-319). In Jack D. Douglas (Ed.) Understanding everyday life. Chicago: Aldine Publishing Company. Bredo, E. (1994). Reconstructing educational psychology: Situated cognition and Deweyian pragmatism. Educational Psychologist, 29 (1), 23-35. Brown, J. S., & Duguid, P. (1993, March). Stolen knowledge. Educational Technology, 10-15. Bruner, J. S. (1983). In search of mind. New York. Harper and Row. Bruner, J. S. (1990). Acts of meaning. Cambridge, MA: Harvard University Press. Clancey, W. J. (1993). Situated action: a neuropsychological interpretation. Response to Vera and Simon. Cognitive Science, 17, 87-116. 28 Clancey, W. J. (1995). Practice cannot be reduced to theory: Knowledge, representations, and change in the workplace. In S. Bagnara, C. Zuccermaglio, & S. Stuckey (Eds.), Organizational learning and technological change. (pp. 16-46). Clancey, W. J. (1997). Situated cognition: On human knowledge and computer representations. Cambridge: Cambridge University Press. Cole, M. (1996). Cultural psychology. Cambridge: Belknap Press of Harvard University Press. Cook, D. A. (1997). Behavioral analysis as a basis for instructional design. In C. R. Dills & A. J. Romiszowski (Eds.), Instructional development paradigms (pp. 215-244). Englewood Cliffs NJ: Educational Technology Publications. Dick, W. (1987). A history of instructional design and its impact on educational psychology. In J. A. Glover, R. R. Ronning (Eds.), Historical foundations of educational psychology (pp. 183-200). New York: Plenum Press. Dreyfus, H. L. (1979). What computers can’t do: The limits of artificial intelligence (2nd ed.). New York: Harper and Row. Forman, E., Minick, N. & Stone, C. A. (Eds.) (1993). Contexts for learning: Sociocultural synamics in children’s development. New York: Oxford University Press. Gagné, R. M. (1965). The conditions of learning (1st ed.). New York: Holt, Rinehart and Winston. Gagné, R. M. (1987). Peaks and valleys of educational psychology: A retrospective view. In J. A. Glover, R. R. Ronning (Eds.), Historical foundations of educational psychology (pp. 395–402). New York: Plenum Press. Gee, J. P. (1992). The social mind: Language, ideology and social practice. New York: Bergen and Garvey. York: Bergen and Garvey. Gee, J.P. (1997). Thinking, learning and reading: The situated sociocultural mind. In D. Kirshner & J. A. Whitson (Eds.), Situated cognition: Social, semiotic and psychological perspectives(pp. 37-55). Mahwah, NJ: Lawrence Erlbaum Associated. Glaser, R. (Ed.). (1965). Teaching machines and programmed learning II: Data and directions. Washington D. C.: Department of Audiovisual Instruction, National Education Association. Greeno, J. G. (1997, January/February). On claims that answer the wrong question. Educational Researcher, 5–17. Greeno, J. G., & the Middle School Mathematics Through Applications Projects Group. (1998). The situativity of knowing, learning, and research. American Psychologist, 53 (1), 5-26. Greeno, J. G., & Moore, J. L. (1993). Situativity and symbols: Response to Vera and Simon. Cognitive Science, 17 (1), 49-59. Greeno, J. G., Smith, D. R. & Moore, J. L. (1993). Transfer of situated learning. In D. K. Detterman & R. J. Sternberg (Eds). Transfer on trial: Intelligence, cognition, and instruction (pp. 99-167). Norwood, NJ: Ablex. Hannafin. M. J. (1997, December). The case for grounded learning systems design: What the literature suggests about effective teaching, learning, and technology. Paper presented at the meeting of the Australian Society for Computers in Learning in 29 Tertiary Education (ASCLTE). [Online]. Available: http://www.curtin.edu.au/conference/ASCILITE97/papers/Hannafink/Hannafink.html [June 15, 1999]. Hannafin, M. J., Hannafin, K. M., Land, S. M., & Oliver, K. (1997). Grounded practice and the design of constructivist learning environments. Educational Technology Research & Development, 45 (3), 101-117. Hilgard, E. R. (1987) Perspectives on educational psychology. In J. A. Glover, R. R. Ronning (Eds.), Historical foundations of educational psychology (pp. 415-423). New York: Plenum Press. Kirshner, D. & Whitson, J. A. (1997). Introduction. In D. Kirshner & J. A. Whitson (Eds.), Situated cognition: Social, semiotic, and psychological perspectives. Mahwah, NJ: Lawrence Erlbaum Associated, Publishers. Kirshner, D., & Whitson, J. A. (1998, November). Obstacles to understanding cognition as situated. Educational Researcher, 22-28. Kliebard, H. M. (1987). The struggle for the American curriculum 1893-1958. New York: Routledge. Labaree, D. F. (1998). Educational researchers: Living with a lesser form of knowledge. Educational Researcher, 4-12. Lave, J. (1988). Cognition in practice. Cambridge, UK: Cambridge University Press. Lave, J. (1991). Situated learning in communities of practice. In L. B. Resnick, J. M. Levine, & S. D. Teasley (Eds). Perspectives on socially shared cognition (pp. 63-82). Washington, DC: American Psychological Association. Lemke, J. L. (1997). Cognition, context, and learning: A social semiotic perspective. In D. Kirshner & J. A. Whitson (Eds.), Situated cognition: Social, semiotic and psychological perspectives (pp. 37-55). Mahwah NJ: Erlbaum. Light, P. & Butterworth, G. (1992). Context and cognition: Ways of learning and knowing. Hillsdale, NJ: Lawrence Erlbaum Associates. Mager, R. F. (1962) Preparing instructional objectives. San Francisco: Fearon. Markle, S. M. (1965). Good frames and bad. New York: John Wiley. Mayer, R. E. (1992). Cognition and instruction: Their historic meeting within educational psychology. Journal of Educational Psychology, 84 (4), 405–412. Mayer, R. E. (1996). Learners as information processors: Legacies and limitations of educational psychology’s second metaphor. Educational Psychologist, 31 (3/4), 151161. Newman, D., Griffin, P., & Cole, M. (1989). The construction zone. Cambridge, UK: Cambridge University Press. Norman, D. A. (1993). Cognition in the head and in the world: An introduction to the special issue on situated action. Cognitive Science, 17 (1), 1-6. Palmer, P. J. (1997, November/December). The heart of a teacher: Identity and integrity in teaching. Change, 15-21. Perkins, D. N. (1991, May). Technology meets constructivism: Do they make a marriage? Educational Technology, 18–23. 30 Resnick, L. B., Säljö, R., Pontecorvo, C., & Burge, B. (Eds.). (1991). Discourse, tools, and reasoning: Essays on situated cognition. Berlin: Springer. Rogoff, B. (1990) Apprenticeship in thinking. New York: Oxford University Press. Salomon, G., & Perkins, D. N. (1998). Individual and social aspects of learning. In P. D. Pearson & A. Iran-Nejad (Eds.), Review of Research in Education, 23, 1-24. Sfard, A. (1998, March). On two metaphors for learning and the dangers of choosing just one. Educational Researcher, 4-13. Shore, B. (1996). Culture in mind: Cognition, culture, and the problem of meaning. New York: Oxford University Press. Shulman, L. S. (1986). Paradigms and research programs in the study of teaching: A contemporary perspective. In M. C. Wittrock (Ed.), Handbook of research on teaching (3rd ed., pp. 3-36). New York: MacMillan. St. Julien, J. (1997). Explaining learning: The research trajectory of situated cognition and the implications for connectionism. In D. Kirshner & J. Whitson (Eds.), Situated cognition: Social, semiotic, and psychological perspectives (pp. 261-280). Mahwah NJ: Erlbaum. Suchman, L. (1993). Response to Vera and Simon’s Situated Action: A Symbolic Interpretation. Cognitive Science, 17 (1), 71-75. Thomas, G. (1997). What’s the use of theory? Harvard Educational Review, 67 (1), 75104. Tripp, S. D. (1993, March). Theories, traditions, and situated learning. Educational Technology, 71-77. Winograd, T., & Flores, F. (1986). Understanding computers and cognition: A new foundation for design. Norwood NJ: Ablex. Young, M. F. (1993). Instructional design for situated learning. Educational Technology Research & Development, 43 (1), 43-58. Author Notes Brent G. Wilson (brent.wilson@ucdenver.edu) is professor of Information and Learning Technologies at the University of Colorado at Denver, with research interests in the adoption and use of learning technologies. Karen Madsen Myers ( karen_myers@ceo.ucdenver.edu ) is a doctoral student at the University of Colorado at Denver, studying professional practices in medical education. Footnotes 1. Jim Greeno has another reason for avoiding the term ‘situated cognition.’ Rather than a kind of cognition, situated cognition is best thought of as the only cognition: "There is not a situated way of teaching and learning that contrasts with nonsituated ways. All teaching and learning are situated; the question is what their situated character is" (Greeno, et al., 1998, p. 19). Clancey (1993, p. 100) echoes this same point: 31 We are always situated because that is how our brains work. We are situated in an empty dark room, we are situated in bed when dreaming. People doing the Tower of Hanoi problem are always situated agents, regardless or how they solve the problem [Situated action] is a characterization of the mechanism, of our embodiment, not a problemsolving strategy. 2. A novel way of looking at the problem of transfer has been offered by Greeno, Smith, and Moore (1993). They frame the issue more ecologically in terms of affordances and constraints that hold across situations. This is in line with the "person plus" unit of analysis, since affordances and constraints cannot be attributed solely to the person or the environment alone, but must be considered in terms of their interrelationship. According to Greeno, when the same kinds of relationships hold across situations, transfer occurs. 3. Adult educators have enthusiastically appropriated the idea of cognitive apprenticeships, but the model has been very unevenly applied in real settings. Whereas Collins and Brown conceived of rich, authentic learning environments, instructional products based on the model can be quite traditional and didactic in nature. Kirshner & Whitson (1998) also note the wide array of methods seeking justification from SitCog theory, ranging from literal apprenticeships to market-driven micro-vouchers as an alternative to public education! 32