Designing an intelligent learning environment (ILE) involves implementing some theory of learning and teaching. However, most available theories do not have the level of operationality required for implementation work.The design of an ILE is a creative process. Our challenge is to develop a framework that builds a bridge between theories and implementation. This framework translates psychological knowledge into terminology more relevant to computer scientists. It specifies the cognitive architecture of MEMOLAB and it is based on two concepts: the pyramid metaphor and the language shift mechanism.
The pyramid is a visual metaphor for the core structure of a learning environment. The pyramid represents the concepts and skills to be acquired by the learner, ranked bottom-up according to their level of "hierarchical integration" (see below). Learning consists in moving up in the pyramid. Each level of the pyramid is defined by two languages: the command language and the description language. The command language vocabulary is the set of elementary actions that the learner is allowed to do at some stage of interaction. The command language syntax defines how the learner composes sequences of elementary actions. The description language is the set of symbols (strings, graphics,...) used by the computer to show the learner some description of his behaviour. This description reifies some abstract features of the learner's behaviour in order to make them explicitly available for metacognitive activities (Collins and Brown, 1988).
The command and description languages are different at each level of the pyramid. The hierarchical nature of a pyramid implies that each level integrates its lower neighbour. This integration is encompassed in the relationship between the languages used at successive levels: if a description language at level L is used as a new command language at level L+1, then the learner is compelled to use explicitly the concepts that have been reified at level L. This is what we called the language shift mechanism (Dillenbourg, 1992): when he receives a new command language, the learner must explicitly use the concepts that were implicit in his behaviour. The meaning of the new commands has been induced at the previous level by associating the learner's behaviour with some representation. This representation is now the new command. The ILE structure can then be described as a sequence of [action language, representation language] pairs, a sequence in which the relationship between two successive pairs is described by the language shift mechanism.
Let us consider a simple example on solving equations. At some level of the pyramid, one can show the learner - with some graphics - that a good heuristic is to regroup the X's on the same side of the equation. At the next higher level, we can offer a button "regroup X" in order to compel him to explicitly use this heuristic in her solution. A more complete example will be described in section IV.3. "Applying design metaphors to MEMOLAB" on page 34.
The role of these metaphors is to translate psychological concepts into design principles. This metaphorical link between theories and design is bidirectional and complex. By bidirectional, we mean that, in one direction, it helps the designer to ground design into a psychological theory, and also that, in the reverse direction, the empirical data collected from using the system constitute a feedback for the theory. This link is complex because it may be used to refer to several theoretical approaches. Most psychological theories address actually only a specific facet of learning while an ILE designer must consider learning in its globality and complexity. For instance, an ILE must account for the importance of discovery, for the role of practice and for the effect of coaching, because all of them occur at some stage of learning in the real world. Therefore, an intermediate framework should integrate multiple theories, each relevant for some aspect of reality. Hare are some examples of links between theories and our design metaphors.
- Piaget (1971) uses the term reflected abstraction for the process by which properties that are implicit at some level of knowledge can be abstracted and explicitly reached at the higher level. The language shift principle translates this idea: the designer first introduces abstract features in the description language, and at then, at the next level, coerces the learner to handle explicitly this features by using them in the command language.
- The pyramid and language shift concepts can be articulated with Vygotsky's idea of apprenticeship. When the learner performs at some level L, the description language describes his activities at level L+1. This level L+1 corresponds to the "zone of proximal development" concept (Vygotstky, 1978). Wertsch (1985) proposed a linguistic analysis of the internalization process that relates it to the language shift. He observed (in mother-child interactions) that moving from the inter-individual to the intra-individual plane was preceded by a language shift inside the inter-individual level: mothers replace a descriptive language by a strategy-oriented language (i.e. a language that refers to objects according to their role in the problem solving strategy).
- During this research we concentrated on one theory: the neo-Piagetian theory of Robbie Case (1985). We focused on this theory because of it's rather operational form, and because, since this theory defines developmental stages, it concerns the macro-structure of Memolab. By macro-structure, we mean the relationship between successive microworlds. Some other theories would be more relevant for the micro-structure, i.e. the interaction among agents inside a microworld.
The key idea in Case's theory is what he calls an "executive control structure". In the eighties, the dominant view in cognitive science was to consider problem solving as the execution of schemata. Case discriminates "figurative schemata" which represent states (from the problem situation until the solution situation) and "operative schemata" which represent transformations. Coherent with the Piagetian tradition, he defines several stages of development: (1) perception of objects and motor activities; (2) relations between motor activities; (3) manipulation of dimensions (quantifiable variables) and (4) second order dimensions (ratios). Within each stage, Case distinguishes sub-stages where the subjects apply increasingly complex schemata. This complexity is defined by the number of "basic units of thought". Each sub-stage is characterized by the subordination of a new basic unit to the executive control structure: the first sub-stage has two basic units, the second has three and the third has four. The most interesting aspect of his contribution concerns the transition between stages. The complexity of schemata reached at the last sub-stage of stage N corresponds to a basic unit at stage N+1. The four-unit control structure of stage N is turned into a one-unit control structure at stage N+1. This process is called "hierarchical integration".
According to Case, the ability to apply increasingly complex schemata is mainly due to the evolution of the subject's working memory (that Case calls "Short Term Storage Space"). From a developmental perspective, working memory increases with age as a result of the maturation of the nervous system. In addition, intensive practice of schemata leads to their automation or compilation, freeing up memory resources for additional basic units. The original contribution of Case is the understanding of why learning is not a simple linear process. It includes quantitative changes within a stage resulting from practice effects, to qualitative changes between stages resulting from knowledge restructuring.
There is an obvious mapping between the structure defined by Case and our intermediate framework. The control structures at each level of the pyramid integrate the control structures located at the lower level. The sequence of microworlds within the pyramid is structured along Case's view of development: quantitative variations define possible improvement within some level (or microworld or stage) while the qualitative variations define the transition between two levels. The concept of stage transition is translated into the language shift mechanism. This transition is necessary when the learner tries to solve problems that have too high memory load constraints. After the language shift, the learner has at his disposal new control structures that enable him to solve the problems with a reduced cognitive load. This ILE-oriented re-expression of Case allowed us to ground the design of MEMOLAB in this theory.
MEMOLAB is structured according to the principles described above. The MEMOLAB pyramid has three levels. Creating an experiment means assembling experimental `building blocks' on the lab workbench. These `blocks' or units are `events' at the first level, `sequences' at the second level and `plans' at the third level. An event (level 1) is the association of some subjects with some task and a measurement device. A task is defined by an activity, a material and some parameters. A sequence (level 2) is the set of activities performed N times by a same group of subjects, with different materials equivalent to the material specified by the learner. A plan (level 3) is a table of factors and modalities that will instantiate a sequence in which some parameters have been replaced by a variable.
Figure 14: The Cognitive Architecture of MEMOLAB
An overview of the learner's activities at each level of the pyramid is presented in Figure 14. For each level, we represent two screens, one displaying the command language and the other presenting the description language. At level one, the novice sequences events on the lab workbench. At the end of level 1, the learner receives challenges that induce the need for comparisons. The concept of sequence is reified at level 1 and used at level 2 as the way of designing new experiments. At the end of the second level, the necessity of taking into account the interaction of effects is induced by new challenges. The concept of plan is presented at level two and used for designing experiments at level 3.
The shift from one level to another, i.e. shifting from one language to another corresponds to some qualitative jump in learning. Within each level, we vary in a more continuous way the difficulty of the challenges proposed by the tutor. The last challenges of level 1 already have the difficulty of level 2. This shows the learner the necessity to have more powerful control structures to solve the proposed challenge. (As in Case theory sub-level i.4 is equivalent to sub-level i+1.0). The difference between levels depends not only on the number of parameters in the experiment but also on more subtle aspects of the task. For instance, at level 2, the described experimental sequence must be considered as the abstraction of something that must be replicated N times. The number of replications varies according to the duration of each experiment in order to compromise between the subject's tiredness and the need for multiple measurement. Another tricky part of the expertise, the selection of interesting parameter values, is brought into the game at level 3.