Two dynamic models of learning based on E. Thorndike's «Cat- in- the box» experiment

In the article it has been shown how the approaches of qualitative theory of dynamic systems may be applied to analyze classical data received by E. Thorndike in his experiments with cats on learning. Two dynamic models have been considered: a well-known Verhulst model and a linear model. Hypotheses as regards the sense of the models' structure have been put forward. Through a numerical approximation of Thorndike's data it has been shown that Verhulst model gives somewhat better results in explanation of the data than the linear model.