Design of Microworlds in Mathematics and Science Education

Microworlds have had a variety of definitions. Weir (1987) stated that the term was used initially by workers in artificial intelligence (AI) to describe a “small, coherent domain of objects and activities implemented in the form of a computer program and corresponding to an interesting part of the real world. Since the real-world counterparts were typically very complex, the microworlds of those early days were simplified versions of reality, acting as experiments to test out theories of intelligent behavior” (p. 12, emphasis in original). For example, researchers at MIT used such environments to develop AI programs that could move blocks around an environment, and learned more from studying how to “teach” computers how to solve such ostensibly simple tasks than they did from studying more complex tasks (Minsky, 1986, see also Papert, this volume). From these roots came the notion that such focused, structured environments might have potential to support learning, and the focus of “microworlds” changed from teaching computers to solve problems in a limited domain to designing educational environments. Most of these environments served as a concrete (c.f., Clements, 1999) instantiation of an aspect of mathematics or science, consistent with diSessa’s (1986) fecund suggestion not to turn learners’ “experience into abstractions [but to turn] abstractions, like the laws of physics, into experience” (p. 212). Papert described a microworld as a self-contained world in which students can “learn to transfer habits of exploration from their personal lives to the formal domain of scientific construction” (Papert, 1980, p. 177). He helped create what is perhaps the most influential example—the turtle geometry component of the Logo computer language, which was conceived broadly as “an object to think with”