A Connectionist Model of the Development of Velocity, Time, and Distance Concepts

Finally, these same-aged children would be expected to solve distance inference problems based on either time or velocity information depending on which is more salient. Therefore, future research should include younger children and manipulate the salience of velocity and time information in the distance inference task. Third, if the task demands of velocity inferences were reduced, it is predicted that 10-year-olds would make velocity inferences by integrating the dimensions with the correct multiplicative rule. Therefore, the inability of Wilkening's subjects to correctly integrate time and distance information is again likely due to extra memory demands. Finally, the simulations suggest that the distance multiplicative stage emerges after both time and velocity multiplicative stages. Since Wilkening did not study 10-year-olds' performance when an eye-movement strategy was not possible, it would be necessary to reexamine 10-year-olds under this condition. In conclusion, the results of our simulations suggest that limited processing capacities of children lead to performance that is first characterized by identity rules for time and velocity inferences. Later, additive inference rules characterize performance. Finally, multiplicative rules characterize the integration of distance, time, and velocity inferences. Once again the cascade-correlation algorithm has demonstrated the ability to capture the emergence of rule-based stages in cognitive development.