A Study of Human Collaborative Discovery Processes of Scientific Concepts.

本研究旨在从认知科学的角度探讨两个人合作发现科学概念的过程。主要结果总结如下：(1)首先，我们基于认知心理学中广泛使用的协议分析法分析了发现过程。在一项研究中，两个人被要求使用专门为这项实验开发的计算机微世界来寻找“遗传法则”。通过观察两个被试之间的互动，我们确定了激活发现过程表现的一些重要因素。(2)其次，我们还讨论了基于计算机模拟方法的发现过程。在一项研究中，我们构建了一个计算模型，可以解决“沃森的2-4-6任务”，这是该研究领域最受欢迎的任务之一。我们仔细控制模型的参数进行了模拟，并对结果进行了分析，发现了一些有效的启发式假设的重要知识。这些发现不能仅仅通过心理学实验来证实。(3)最后，我们提出了开发支持科学概念学习的CAI系统的三个阶段（准备阶段、开发阶段、评估阶段）。为了验证这个想法，我们实际开发了一个通过认知地图和心智模型学习初级电路的CAI系统，并证实了这个想法的实用性。

The purpose of this study is to investigate the collaborative discovery processes of scientific concepts by two persons from view point of cognitive science. The primary results are summarized as follows.(1) First, we analyzed the discovery processes based on the protocol analysis method that had been widely used in cognitive psychology. In one study, two persons were asked to find "the laws of heredity", using the computer microworld that was developed for this experiment. Through observing the interaction between the two subjects, we identified some important factors that activated the performance of the discovery processes. (2) Second, we also discussed the discovery processes based on the computer simulation method. In one study, we constructed a computational model that could solve "the Wason's 2-4-6 task" that was one of the most popular tasks in this research domain. We executed the simulations controlling carefully the parameters of the model, analyzed the results, and found some important knowledge of effective heuristics for hypothesistesting. Those findings could not be recognized merely by psychological experiments. (3) Last, we proposed the three stages (preparation stage, development stage, estimation stage) for developing CAI systems that supported learning scientific concepts. For testing this idea, we actually developed a CAI system for learning elementary electrical circuits by cognitive maps and mental models, and confirmed the utility of this idea.