Jerry Willis
Contributing Editor

In the educational technology literature the term instructional design has so many meanings that its use has little purpose without further elaboration. In some publications, the term roughly refers to the field called educational technology in this article (e.g., Riegeluth, 1983). In others, the terms is used to describe the practice of educational technology from a particular theoretical perspective such as behaviorism (Gropper, 1987). Instructional design may also be the umbrella term indicating the components that should be included in an instructional package (Merril, 1988). The phrase has also been used to indicate the process others have used it to mean a particular model or theory that can guide the design of instruction (Wright & Conroy, 1988). In this article, instructional design (ID) refers to the process of designing materials, and the term instructional design model (ID model) refers to a model or theory that can guide the instructional design process.

Over the years many instructional design models have been proposed (Bagdonis & Salisbury, 1994). However, the great majority of the models currently available are based on social science theories from the behavioral family, broadly defined to include information processing and cognitive science theories that "break down" content to be taught into smaller units which are then taught with direct instruction strategies. Dick (1995), for example, comments that the "historical roots of what today is referred to as instructional design was Skinnerian psychology, especially as it was manifested in programmed instruction" (p.5). To the Skinnerian base, Dick adds three other influences: the instructional objectives of Robert Mager, Robert Glaser's criterion-referenced testing, and Robert Gagne's events of instruction. All these influences, from Skinner to Glaser, fall well within the behavioral family of theories, which it is suggested here, includes cognitive science and information processing theories that share a common foundation with more extreme versions of the theory such as radical behaviorism. The term "behavioral" will be used in this article to indicate this family of theories, but other terms are also appropriate, including objectivist, empirical, and rational. Winograd and Flores (1987) called this group the "rationalistic tradition," while Schon (1987) used "technical-rational" in much the same way. Many of the traditional theories of learning from psychology fall into this family as well as many of the instructional theories developed by educational technologists. The Gagne-Briggs (Gagne & Briggs, 1979) and Gropper's (1987) behavioral approach to instructional design come immediately to mind.

While there are real differences within the behavioral (or objectivist or rationalistic or technical/rational) family of theories, they all share related philosophies of science that are objectivist, rational, and empiricist or postempiricist. And the tenets of those philosophies of science can be seen in the instructional design models derived from them. This paper lays the foundation for an alternative instructional design model based on social science theories from the constructivist family and on an interpretivist philosophy of science. These two families of theories-behavioral/objectivist/empiricist/rational and constructivist/interpretivist - have radically different answers to two fundamental questions, and those answers are the reason for many of the differences between a behavioral-empirical instructional design model and a constructive-interpretive ID model. The two questions involve the role of language and the nature of truth.

Objectivist-rational approaches consider language a theory-neutral medium through which meaning about an external world can pass without being influenced or changed. To fulfill this role, language must therefore be clear-cut, precise, and unambiguous. The ability to use such a language is an essential foundation for many research and development traditions in both psychology and computer science. In psychology, information processing theory as well as much of the work in cognitive science is based on this perspective of language. In computer science, efforts to create such a language was an important part of Alan Turing's work as well as of those who followed Turing's trailblazing path. Newell and Simon's general theory of problem-solving is based on this theory of language (Simon, 1991) and is still very influential today. Much of the computer-based research in cognitive science as well as research on artificial intelligence and expert systems that is in the Newell and Simon tradition is also based on the behavioral, objectivist-rational model of language. In fact, the assumption that the precise, crisp way a computer interprets a program could be used as a model for the way humans use language is at the heart of many research programs over the last thirty years. In educational technology, the emphasis on creating and using an objective language is reflected in the emphasis on creating instructional objectives during the 1960s-1980s.

The purpose of a specific objective is to describe precisely what the student will be able to do after completing a segment, lesson, or unit of instruction. Using the specific objectives, we write text, interactions, and test items. A specific objective includes (1) the conditions under which the student will perform, (2) the action required, (3) how the student will demonstrate the action, and (4) the mastery level required. (Criswell, 1989, p. 58)

Perhaps the best-known component of the instructional design model is the writing of performance objectives, or, as they are more commonly called, behavioral objectives. (Dick & Carey, 1985, p. 97)

While many theories used in computer science, educational psychology, and instructional science take the theory of language discussed here as a given, there are many critics who find both the theory and the results of the theory less than satisfying. The book by Winograd and Flores (1987), for example, is a cogent argument against the behavioral model of language in computer science, and it is not surprising that the authors question many of the fruits of this tradition, including the Newell and Simon general theory of problem solving, artificial intelligence, and expert systems.

Criticism of the behavioral role of language has been made for over a hundred years. In this century, one especially influential critic was an Austrian philosopher, Ludwig Wittgenstein. in fact, Wittgenstein and Alan Turing, one of the founders of modern computer science, debated the role of language in both human and machine thinking in 1939 (Diamond, 1976). The Wittgenstein-Turing debates highlight one of the fundamental points of disagreement that continues to influence both research and practice in the field of educational technology today. For Turing, and others in the behavioral or objectivist-rational tradition, language is a means of describing the real world. When properly used, it is a passive vessel for communicating that does nothing more than transfer information.

In the behavioral/objectivist, rational tradition, which includes empiricists and postempiricists, it is essential that language play a neutral role n the knowledge transmission process because the theory of knowledge they use requires several types of separation - separation of the knower from what we wish to know, separation of fact from value, separation of objective data from subjective data, separation of what you "want" reality to be from what the data shows it to be (Smith, 1989, 1993). Since language is the medium of knowledge transmission, if it cannot be made to be "neutral,' then dichotomies such as those noted above collapse. If language itself helps to determine reality, or influences how we define it, then there is no possibility of a purely objective or rational approach. For objectivists, then, the crucial aspect of language is its relationship to an external reality.

The language-external reality connection, however, is not the critical relationship for Wittgenstein. He concluded that efforts to explain language as a logical reflection of an "out there" reality were futile. He proposed, instead, the language game concept. Games, in fact, are central to several of Wittgenstein's positions. He argued that a language is like a game. You learn the rules of a game by playing the game, not by studying a rulebook. Similarly, you learn the game of language by playing that game. Language acquires meaning through use, not study of a dictionary.

For not only do we not think of the rules of usage - of definitions, etc. - while using language, but when we are asked to give such rules, in most cases we aren't able to do so. We are unable clearly to circumscribe the concepts we use; not because we don't know their real definition, but because there is no real "definition' to them. To suppose that there must be would be like supposing that whenever children play with a ball they play a game according to strict rules. (Wittgenstein, 1965, p. 25)

Wittgenstein's explanation of the role of language calls into question several lines of research in cognitive science-such as using computers running a program to model the way humans use language-as well as common practices in educational technology, such as writing precise, detailed instructional objectives early in the instructional design process that then guide later work. if meaning derives from use, as Wittgenstein suggests, then writing detailed objectives early in the design sequence ignores the fact that meaning and understanding are likely to develop and emerge throughout the process. This point has been acknowledged by several educational technologists, including some who still use a behavioral ID model.

Some instructional designers insist that learning objectives be stated carefully early in planning.... Sequentially, this may sound correct, but in actual practice it does not often work. While some instructors can write their learning objectives immediately, many

people cannot enumerate detailed objectives at this point. The reason is that their own thinking about what to include in the unit may not yet be clear....

Writing learning objectives is a developmental activity that requires changes, refinements, and additions as the writer develops subsequent planning steps. Sometimes it is not until learning activities are being selected or evaluation methods stated that the 'real' objectives for teaching a topic become clear. Therefore, expect to start with loosely worded objectives, move ahead in the planning sequence, and then return to spell out the learning objectives in specific detail as each one becomes evident. (Kemp, 1985, p. 78)

Another common ID strategy, the use of a designer who is external to and unfamiliar with the students or the context of practice to prescriptively create instruction also becomes very questionable. Wittgenstein's view of language would support a more contextual approach to ID that puts the designer in the environment of use for an extended period of time in order to understand that context. Giacomi, Mosher, and Swenton-Wall (1993), for example, suggest taking an ethnographic field methods approach to design. While their focus is primarily on the design of products and software for use in organizations, the points they make are also applicable to instructional design:

The ethnographic approach, with its emphasis on "natives' point-of-view,' holism, and natural settings, provides a unique perspective to bring to bear on understanding users' work activities...

This emphasis on natural settings derives in part from a belief that particular behaviors can only be understood in the everyday context in which they occur. To remove a behavior from the larger social context is to change it in important, nontrivial ways...

Ethnographic field work requires the personal involvement of the investigator, a willingness to be in situations out of one's control and as such an abandonment of strict 'scientific control.' It also involves an iterative, improvisational approach to understanding, wherein partial and tentative formulations are revised as new observations challenge the old, and where adjustments in research strategy are made as more is learned about the particular situation at hand. (Giacomi, Mosher, & Swenton-Wall, 1993, 123-130)

Once he had at least partially disconnected language from an external reality, Wittgenstein went one step further. When striving to understand a concept, traditional logic calls for definitions that prescribe the essential elements or characteristics of something. An apple, for example, has much in common with a pear, but distinguishing between them could be approached by creating a list of characteristics of all apples that are not characteristics of pears. Wittgenstein found this approach less than satisfactory. "According to Wittgenstein, this striving to formulate the 'essence' of a thing - that is, common characteristics deemed to be necessary and sufficient for its existence-has led us astray again and again" (Schulte, 1992, p. 111).

Wittgenstein proposed instead the idea of family resemblance. Members of the same family are often recognizable as kin even though no single feature is the same across all family members. He illustrated his point in a famous comment about the meaning of the word "game."

Consider, for example, the proceedings we call "games." I mean board games, card games, ball games, athletic games, etc. What is common to all of these? Don't say: "They have to have something in common or they would not be called 'games' - but rather look and see if there is something common to all of them. For if you look at them, you will not see something common to all; you will see similarities and relationships-a whole series of them. As I said: don't think, look! - Look, for example, at the board games with their variety of related features. Now go on to the card games: here you will find many things that correspond to the first group, but many common traits disappear while others appear for the first time. If we move on to ball games, much that is in common is retained, but much is lost..... And we can proceed through the many, many other groups of games in the same way, seeing similarities surface and disappear.

And the result of this investigation is like this: We see a complicated network of similarities intersecting and overlapping one another-similarities large and small. I can give no better characterization of these similarities than 'family resemblance' for it is in just this way that the various resemblances to be found among members of a family overlap and intersect: build, facial features, eye color, gait, temperament, etc., etc.-And I would say: 'games' for a family. (Wittgenstein, 1958, 66f)

Wittgenstein thus moves well away from the view of language as something to be improved to the point that it is an objective, transparent medium for communicating information about reality (language as a calculus)an important goal for many theoretical traditions in psychology and educational technology, but something Wittgenstein thought both impossible and less than useful. Language for Wittgenstein is a living, changing entity whose meaning develops through use.

To summarize, two perspectives on language are competing foundational "givens" for both practice and research in educational technology today. Wittgenstein is often cited in the constructivist literature of computer science, and rarely mentioned in the behaviorist literature. Using the same rough division of educational technology theories, Wittgenstein is rarely mentioned in the behavioral literature but regularly cited in papers based on constructivist theory. Wittgenstein's views on language and meaning, for example, are cited as major influences in the development of cognitive flexibility theory (Spiro, Feltovich, Jacobson, & Coulson, 1992a, b) and as a foundation for the situated cognition work of Vanderbilt's Cognition and Technology Group (1990, 1991, 1992); the adoption of an interpretivist as opposed to an empirical approach in educational research (Howe, 1988); and the creation of instructional systems technology as a 'linking science' (Bednar, Cunningham, Duffy, & Perry, 1992). In addition, Richard Rorty, the most influential living philosopher in the pragmatic tradition of John Dewey and William James, uses Wittgenstein as one of the foundations for his democratic pragmatism (Rorty, 1991).

In the area of instructional design, many aspects of behavioral ID models assume that the language-external reality link is critical and that careful use of language allows us to approach the design process in an objective, rational manner. In contrast, Wittgenstein's view of human language, which underpins constructivist approaches to educational technology, undermines several aspects of behavioral ID models and suggests alternative ways of approaching ID. Some of those alternatives will be explored later in this article. The next topic, however, involves answers to a second fundamental question.

There are several approaches to the concept of truth and how it is discovered. Foundationalist approaches such as objectivism and empiricism use an 'out there" approach that assumes we can study a real world and thus compare what we believe to be true with what "really is true" in the external or real world. An external world that is separate from, and independent of, human consciousness thus serves as a foundation for any claims to truth-from the claim that students learn to write better when they use word processors to the assertion that computer-assisted learning is effective and should be used more often in schools.

Asserting that there is a real world out there, however, is not the major point of disagreement. These approaches incorporate a rational tenet which proposes that through the thoughtful use of proper methods such as scientific research-humans can come to know what that external reality is. Experimentation and systematic observation of the "real world' are thus the source of knowledge about reality in the foundationalist approaches such as objectivism and empiricism. For empiricists, reality is the world around us that we can sense. Thus, "sense data," such as the results of research studies, are a primary way of coming to understand reality. Criswell (1989) expresses this perspective in the introduction to her instructional design book when she says her book "emphasizes knowledge from psychology and educational research that is fundamental to CBI design and evaluation. it also describes a systematic design and evaluation process that produces successful courseware if this knowledge is used" (p. v). The foundationalist approach to defining reality is another essential basis for the behavioral approach in educational technology.

A second, and quite different, answer to the question "what is reality?' is the foundation for a social constructivist approach to educational technology that is based on the theories of Vygotsky (1978), the philosophy of Wittgenstein, and the democratic pragmatism of Richard Rorty. In this section the views of Rorty (1982, 1991) will be emphasized because they reflect the current perspective on reality from an interpretivist viewpoint. Central to much of Rorty's philosophy is his definition of objective and subjective. In empiricist and postempiricist philosophies of science, "objective' refers to statements that can be empirically demonstrated to correspond to an "out there" reality, while subjective statements are influenced by background, experience, opinion, values, morals, feelings, and other non-empirical influences. The important relationship, as with the objectivist view of language, is between language and external reality. Rorty (1979) begins his analysis from a very different perspective. He argues that we can never look at any issue "objectively" in the empiricist's sense of the word because we are all influenced by our own culture and experience. Everything is thus "subjective" if the empiricist's definition is used. Rather than abandon the objective-subjective dichotomy, however, Rorty redefines the two terms. He uses the word objective to describe things a group agrees on, while subjective describes things the group does not agree on. In the empiricist framework, objective and subjective are core terms within a correspondence theory of truth (corresponding to external reality), but in Rorty's interpretivist framework they are core terms in a consensus theory of truth (a group comes to consensus). Truth, and reality, are universal in the empiricist theory; they are local and transitory in interpretivist theory.

Much of the work in educational technology that is conducted on an empiricist or postempiricist foundation is based on the core assumptions that objective knowledge is universal knowledge, that objective can be distinguished from subjective and that objective is always better. As noted earlier, many instructional design models, for example, call for the creation of specific, detailed instructional objectives before developing instruction. Objectives (and objective assessment) which are supposed to support and enhance instruction often seem to overwhelm it. In the 1985 edition of their book, The Systematic Design of Instruction, Dick and Carey devoted nine chapters to goals, subgoals, objectives, and assessment (entry behaviors, exit behaviors, grading, formative evaluation, and summative evaluation) and only three to the creation of instructional materials. Some (including this author) would argue this is the tail wagging the dog.

The objective approaches are based on a value-based dichotomy-objective versus subjective-that has been collapsed by Rorty and other interpretivists. Interpretivists deny the possibility that humans can take a "God's-eye view" and make objective decisions. Objective, in the empiricist sense, simply does not exist within Rorty's epistemology. Even if it is desirable and preferable, it cannot be attained by humans. if the interpretivist perspective is accepted, many commonly accepted aspects of educational technology practice come into question. Can we justify heavy reliance on behavioral objectives to guide both instructional design and teaching? Are the many hours of instructional time spent on objective assessment of learning objectives worthwhile? Does it make sense to follow the traditional linear sequence of activities in an instructional design model that begins with learner, task, and concept analysis, followed by design, development, and summative evaluation? All these activities are based, at least in part, on an assumption that objective can be distinguished from subjective and that objective is better.

lnterpretivists are antifoundationalists, they believe "there is no particular right or correct path to knowledge, no special method that automatically leads to intellectual progress" (Smith, 1993, p. 120), a position that is particularly troubling to empiricists and postempiricists, who believe no progress can be made unless "standards' are upheld. Interpretivists don't have a problem with standards-in research, in the process of designing instructional materials, or in professional practice. it is just that they do not believe those standards are in any way universal or special. They are, instead, the products of a particular group or culture. Interpretivists don't abandon standards, they simply accept that whatever standards are used are subjective, and therefore fallible, rather than objective and universal.

Because the differences between foundational and antifoundational approaches to reality are so profound, and have so many implications, many question whether there can be a rapprochement or interaction between educational technologists on different sides of this issue. The foundational versus antifoundational question is at a much different level than decisions on whether simulations or tutorials are preferred for a particular instructional project. Dick (1995) is optimistic about the compatibility of the two approaches and is critical of those who are not:

Constructivists have argued that you are either with them or against them. You can't accept only some of their ideas and reject the rest (Bednar, Cunningham, Duffy, & Perry, 1 992). I disagree With this position… I would argue that designers who augment ISD fundamentals with judicious use of selected Constructivist principles will make design decisions that result in instruction that is both engaging to learners and produces learning outcomes that are required by the client. What is required is a balanced perspective, and a balanced set of criteria by which we evaluate our efforts. (Dick, 1995, p. 10)

Dick seems to be proposing a reasonable and "balanced" approach as an alternative to Bednar, Cunningham, Duffy, and Perry (1992), who argue that:

... abstracting concepts and strategies from the theoretical position that spawned them strips them of their meaning. Theoretical concepts emerge in the context of certain epistemological assumptions which underlie the theory. To use a concept like knowledge of results stripped from the assumption that learning is the strengthening of S-R bonds strips the concept of its fundamental basis. (p. 19)

But Dick's approach is not balanced; it proposes adding a bit of constructivist seasoning to the behavioral ID stew. Even when you add the seasoning you are still eating behavioral stew. When Dick suggests designers "augment ISD fundamentals" with "judicious' bits of constructivism, he seems to be acknowledging what the constructivists have been saying for some time-it is the fundamentals, the bedrock assumptions and conceptual framework within which we operate, that are most important. Constructivists are not simply arguing that we should begin to use a few more instructional strategies with names like anchored instruction, they are arguing that we ought to look at learning and teaching from a different framework, one that begins with a different set of givens. In our book on constructivist approaches to using technology in reading, language arts, and literacy instruction (Willis, Stephens, & Matthew, in press), we make a similar argument from the other side: adopt a constructivist approach to instruction, including the foundational assumptions of the approach, and then judiciously use direct instruction (e.g., behavioral) strategies such as mini-lessons and tutorials, just-in-time instruction, and other types of teacher-centered methods to support student-centered learning. We do not, however, believe that we are adopting a behavioral approach or even combining the two approaches. The foundation of our approach remains constructivist. The same is true of Dick's approach; it remains behavioral. While he will accept the idea of including some constructivist instructional strategies, his approach remains firmly anchored in the foundationalist, empirical-rational, behavioral tradition.

That is made clear in his comments about constructivist approaches to instruction:

The constructions of our critics will likely find a place in our education and training programs, but I am not sure why. They often assume entry behaviors that learners don't have. They provide practice on skills that have never been mastered or were mastered long ago. They purport to encourage problem solving but offer little evidence of such because of the constructivist aversion to criterion-referenced testing, that is, the reluctance to establish standards of performance that all learners will achieve. The "assessments' that are utilized are usually in the form of presentations of portfolios of completed activities. No value judgments are made. But, is it educationally worthwhile if the learner is simply engaged, and has fun? (Dick, 1995, p. 10)

That these comments are polemic is probably best demonstrated by asking knowledgeable educational technologists from each camp to read the passage and then comment. Constructivists are likely to sputter a bit and then have difficulty deciding where to start because there are so many fundamental points on which to disagree and so many disagreeable but unspoken givens that are implicit in the statements. Dedicated behaviorists, on the other hand, are likely to have a warm, fuzzy feeling about these comments. They just seem to "feel" right because the unspoken givens and the positions taken on fundamental points are "in sync" with the behavioral perspective.

Some philosophers of science argue that different theoretical perspectives are incommensurable, which is to say there is no way to develop a way of testing the validity of different theories that members of different theoretical camps will agree is a satisfactory test. Behavioral and constructivist ID models may be incommensurable in both the philosophical and practical sense. There is often, however, a pattern to the process of responding to "incommensurable" theories and models. In his discussion of the critical response to pragmatism, William James, in 1907, commented:

I believe that our contemporary pragmatists, especially Messrs. Schiller and Dewey, have given the only tenable account of this subject. It is a very ticklish subject, sending subtle rootlets into all kinds of crannies, and hard to treat in the sketchy way .... But the Schiller-Dewey view of truth has been so ferociously attacked by rationalistic philosophers, and so abominably misunderstood, that here, if anywhere, is the point where a clear and simple statement should be made.

I fully expect to see the pragmatist view of truth run through the classic stages of a theory's career. First, you know, a new theory is attacked as absurd; then it is admitted to be true, but obvious and insignificant; finally it is seen to be so important that its adversaries claim that they themselves discovered it. (James, 1907, quote from Kuklick, 1981, p. 25).

Many of the disagreements on fundamentals that -separated rationalists and pragmatists in the early 1900s are with us today. At the level of professional practice, many of them are at the heart of differences between behavioral and constructivist approaches to instructional design. It may be quite instructive to keep in mind James' description of the "classic stages of a theory's career" while reading behavioral criticisms of constructivist theory.

Although there are a great many instructional design models, many, perhaps a majority, come from the behavioral, objectivist-rational tradition rather than the constructivist-interpretivist tradition. The various behavioral, objectivist-rational ID models currently available differ in many ways, but they do exhibit many family resemblances. Most of them-including the 4D model of Thiagarajan, Semmel, and Semmel (1974), Dick and Carey's (1985) Systematic Design Model, and Criswell's (1989) model for creating computer-based instruction-take a similar approach to common issues. Some of the family characteristics of behavioral ID models, especially those used to create computer-supported learning, are listed in Table 1.

The eight positions listed in Table 1 reflect much of the "core" of objective-rational thinking on procedural instructional design models. Procedural ID models "describe how to perform a task and are formulated to simplify and explain a series of complex processes" (Bagdonis & Salisbury, 1994). Many of these positions are, in fact, taken as typical of most or all ID models by some:

Procedural models in the instructional systems design (ISD) field... attempt to account for all relevant components using a systematic approach to designing instruction, from needs assessment through the development of material, implementation, and evaluation. Each component within the process builds upon the other....

Traditional ISD models are generally viewed by individuals in the instructional design field as representing a linear process, a plan of separate steps that proceed in a linear sequence.

A typical ISD model is divided into five stages: analysis, design, production/development, implementation, and maintenance/revision .... The five stages consist of an integrated set of components that are sequenced so that each component within the process must be completed before continuing to the next. (Bagdonis & Salisbury, 1994)

CBI [Computer-based instruction] applications are developed from a common basis..... Skinner (1984) suggests four components of successful instruction (in any application) in the functional or behavioral tradition:

At least one survey of what instructional designers actually do also reflects a behavioral-empiricist ID model. Wedman and Tessmer (1 992) found that 94% of their sample said they "always" or "usually" wrote learning objectives, for example. The percentage for summative evaluation was 75%. But when asked about formative evaluation, only 49% said they always or usually included it in their work.

Although the behavioral, objective-rational approach is well entrenched in both practice and in many of the textbooks on instructional design, an alternative to this approach has emerged over the last decade. While constructivist approaches have not replaced behavioral approaches as the dominant theoretical framework, they have made significant inroads as evidenced by special issues of Educational Technology on the topic (May & September, 1991; March, 1993; February, 1994; October, 1994). Most of the literature on constructivist approaches to educational technology focus, however, on instructional theory rather than instructional design models. There is a healthy literature, for example, on anchored instruction, situated cognition, and cognitive flexibility hypertext. Very little, however, has been written on the instructional design process itself. There are many possible constructive-interpretive ID models, but they all might share perspectives that resemble many of those in Table 2.

While the positions in Table 2 are derived from the constructive-interpretive literature, it is interesting to compare them to suggestions made by a one of the creators of the leading behavioral ID model. Walter Dick (1 995), in a response to critics who are members of what he called the 'Situated Cognition/Constructivist/Anchored Instruction crowd," suggests six ways to deal with the criticism that instruction created using behavioral ID models is boring and uncreative. In addition to the standard reply that boring instruction might be due to inappropriate use of the ID mode, Dick offered the suggestions below. Several of his suggestions are related to one or more of the positions in Table 2:

"Extensive formative evaluation" (p.7). "if 'creative instruction' is an important characteristic.... then an array of questions can be asked during the formative evaluation to determine how creative the instruction is perceived to be by the learners' (p. 6). (This seems to support positions I and 6 in Table 2.)

"Broad subject matter knowledge, or access to it" (p.7). (Here Dick recognizes a major issue; see position 4 in Table 2.)

"Extensive use of learner analysis" (p. 7). Learn more about the target population "and match their interests and concerns in the instruction" (P. 6). (This and Dick's next two suggestions seem to support the thrust of position 5 in Table 2.)

"Knowledge of the context in which the newly-learned skills will be used" (p. 7).

Use "various instructional strategies" (p. 7). "In the Dick and Carey model of instructional design.... it is suggested that designers provide certain introductory information, along with information and practice and feedback for each of the major objectives to be taught. There is a great deal of freedom here in terms of how information will be clumped, and what kinds of practice activities will be used. There is no reason why other theoretical approaches could not be used to formulate the strategy for conveying information to the learner" (p. 6).

Table 2. Family Characteristics of Constructivist-interpretivist Instructional Design (ID) Models.

Dick thus addresses four (see numbers 1, 4, 5, and 6 in Table 2) of the seven positions I suggest characterize constructive-interpretive ID models. While this is encouraging, the overlap is much less than it appears to be. For example, Dick promotes using formative evaluation to determine whether students perceive the instruction as creative. A constructivist might actually involve students in the design process and co-author material with students and teachers.

Dick also points out that 'the greater the knowledge the designer has about both his or her own area of expertise and the area in which the instruction is being developed, the greater the likelihood that innovative approaches can and will be taken in the instruction" (p. 7). However, while the constructive-interpretivist approach asserts that general ID experts don't exist, Dick concludes that "Since designers are often working in subject-matter domains in which they have no formal training, they must rely on subject-matter specialists to provide the content information. This can be a constraint on the creativity of any designer' (p. 7). Blaming the teacher, or subject-matter expert, does not get us very far. There are many alternatives to simply accepting the limitations of a designer who is external to and unfamiliar with both the context of instruction and the content. Traditional ID models accept the limitation because they are based on the assumption that the ID expertise of the designer will outweigh the lack of understanding about subject matter and context. Even if that were true, there are alternatives. Participatory design models (Schuler & Namioka, 1 993), for example, immerse designers in the context of design and, through collaborative design efforts, help them better understand both context and content. Or, if the designer is willing to give up control to others, he or she might take the role of consultant and facilitator to others, such as subject matter experts, who actually design the instructional material.

Dick's discussion of instructional design models also accepts as givens, implicitly or explicitly, most of the positions in Table 1. For example, he discusses the need to "analyze the instructional goal in order to identify the subordinate skills that must be included in the instruction' (p. 7) and comments that "The process is used to identify exactly those skills that are required by the learner to master the terminal objective, and, by inference, to reject skills and knowledge that are not required to achieve mastery. Designers are always admonished to eliminate the 'nice to know' so that the instruction will be efficient as well as effective" (p. 7). This approach seems to strive to eliminate incidental learning, 'teachable moment" opportunities, and the "target rich" learning environment proposed by some theories. Also eliminated are many of the opportunities for students to make decisions about their learning that are also encouraged by some theories. it is difficult to see how Dick's directions would be applied to an instructional approach such as anchored instruction, or any other approach in which at least some of the objectives and/or learning strategies are developed, created, or selected by the learners themselves. It does, however, reflect a model of instruction in which the objectives of instruction are predetermined by experts, the content is broken down into subskills that are taught directly, and context is not considered critical.

Much of Dick's paper is, in fact, organized around two pseudo issues-that critics say traditionally designed instruction is boring and the solution suggested is more "creative' instruction. However, the terms "boring" and "creative" are sufficiently fuzzy and ill defined to make them relatively useless in a discussion of the critical issues that separate behavioral, objective-rationalist ID models from constructive-interpretivist models. Laying aside the fundamental or philosophical issues, at the professional practice level questions like the ones below are more critical:

Do we begin the ID process by creating a set of goals and objectives, or do they emerge in the process of instructional design?

Is ID a linear, sequential process, or is it iterative and non-linear? Or both?

Is ID expertise a stand-alone skill or is it contextual?

Are the instructional approaches associated with terms like authentic, student centered, problembased, situated, and anchored critical to meaningful learning, or do the approaches associated with behavioral models of instruction produce the desired effect? Are direct-instruction approaches best as support strategies for student-centered learning, or should they be the first line of instruction?

To what extent should ID be a collaborative, participatory process versus a top-down process?

Is the collection of significant amounts of objective data critical in both design and instruction, or are more qualitative types of data often more useful?

Dick (1995) deals directly with only one of these issues-that of ID as a linear process. He describes the criticisms of the strict linearity of the standard ID model as "a straw man of the first magnitude" and expresses surprise that this argument is taken seriously by anyone. I in) afraid I am one of those ignorant critics who, in spite of Dick's efforts to explain the situation, still see much that is linear in the Dick and Carey ID model as well as most of the others in the behavioral, objective-rational tradition. Many steps in the model, for example, are based on previous steps which makes it difficult to complete the process in a non-linear fashion. in addition, Dick and Carey's (1 985) notes to instructors include the comment that "The text presents the model components in the sequence typically followed when designing instruction" [italics added] (p. viii-ix).

In criticizing the behavioral ID model, constructivists have a target-rich environment, with the Dick and Carey model being the best known. On the other hand, the seven positions listed in Table 2 do not, collectively, constitute an ID model. However, any ID model based on constructivist-interpretivist theory should be compatible with most of the positions In Table 2. One such model is the R2D2 model.

[part 1]     [part 2]      [back to module]

Willis, J. (1995). A recursive, reflective instructional design model based on constructivist-interpretivist theory. Educational Technology, 35 (6), 5-23.

© 1995, Educational Technology
Permission being sought.