Perception of Number

Abstract

Configurations varying in arrangement of elements were constructed to assess whether number is abstracted in estimation tasks. In some configurations, circles (or squares) were placed in separate groups of two or four, and interspersed randomly among single randomly placed squares (or circles). Other configurations consisted of single circles and squares randomly placed. Besides arrangement, the proportion of circles or squares varied from 0.40 to 0.60. It was expected that estimations would systematically favour the category that was arranged in separate groups. This effect was found in all experiments, but was not larger for group 4 compared to group 2 arrangement. Moreover, the effect occurred for eight year-old and older subjects, and could be induced in six year-olds who did not show the effect prior to training by training them to abstract number. Results were interpreted in terms of the perception of a higher order structure of number.