Unlearning Aristotelian physics: a study of knowledge-based learning
A study of a group of elementary school students learning to control a computer-implemented Newtonian object reveals a surprisingly uniform and detailed collection of strategies, at the core of which is a robust "Aristotelian" expectation that things should move in the direction they are last pushed. A protocol of an undergraduate dealing with the same situation shows a large overlap with the set of strategies used by the elementary school children and thus a marked lack of influence of
... m physics training on this student's naive physics. The data from these two studies are pooled and elaborated into a "genetic task analysis" of how one might come to understand Newtonian dynamics as a more or less natural evolution from the naive state. 42 oISESSA the right." In contrast, the Newtonian dynaturtle moving upward has momentum in the upward direction which is not affected by the sideways kick, and thus it takes a "compromise" path away from the kick as shown in Figure 3c . The vector addition relevant to this is that shown in Figure 1 . All of the students spontaneously generated the sideways kick as a means of making a right turn, and expressed surprise and consternation at the result. Complaints that the machine was not working correctly at this point were commonplace and vociferously made. The robustness of the students' theory is attested to by the fact that, though many of the students had made significant progress in the two weeks of exposure, none proved to completely shed the Aristotelian disposition. Equally intriguing and more important from the standpoint of making a genetic task analysis is that despite an incorrect "theory," students proved capable of developing alternate strategies for dealing with the corner situation, strategies based on other ideas more compatible with Newtonian dynamics. In fact, several achieved practical mastery of the dynaturtle in most circumstances and proceeded to use it in , projects of their own.