Abstract

This study was planned to investigate whether structural alignments have an influence on categorization processes through comparing the effect of structural alignment on similarity-based transfer and rule-based learning in categorization. In Experiment 1, I manipulated common attributes and dimensional correspondence between learned exemplars and new exemplars. I found that categorization performances in general declined as the number of common attributes decreased, but this tendency was statistically significant only in the match in place condition. These results implicated that similarities between learned exemplars and new exemplars had an influence on categorization and the influence of similarities grew larger as stimulus's attributes had been well aligned. In Experiment 2, I instructed participants to learn categories by rules and examined whether structural alignment had an effect on the rule-based categorization by comparing the match in place condition with the match out of place condition. I found that participants learned categories more quickly in the match in place condition than in the match out of places condition. In conclusion, structural alignment between stimuli play an important role in categorization processes.