The Difficulties In The Categorization Tasks Effect On Two Kinds Of Category Learning

On the base of the former studies about category learning, Author investigated that the difficulties of categorization tasks were how to affect the rule-based and information-integration category learning for the issue of denying the effects in the multiple-systems theory researches. The difficulties of classification tasks are expressed by discontinuous category distribution and adding relevant dimensions and category number. There are three researches in our studies include that the relevant dimension number effects on the category learning, the discontinuous categories and the category number impact on the category learning.Relevant dimension number effects on category learning were investigated in research 1. The differences between rule-based and information-integration category learning were supposed as unequal relevant dimension number which was represented. If the two categories learning had equal relevant dimensions represented and difficulties, the differences would disappear. In addition to comparing the two category learning, two rule-based 2 dimensions category structures were designed as comparing with information-integration category structure. The results in research 1 showed that the learning in rule-based unidimensional structure was better than that in information-integration, and that no significant differences was found between rule-based 2 dimensions continuous structure category learning and information–integration category learning, which was better than rule-based 2 dimensions discontinuous structure.The subjects'response strategies were also approached in research 1. The results showed that the response strategies were influenced by rule complexity. The most subjects could use optimal rule in unidimensional rule-based structure tasks, and the majority could use rule-based strategies whereas the minority could use similarity strategies in information-integration structure tasks. The majority could use optimal strategies whereas the minority could use suboptimal rule-based strategies in 2 dimensions continuous rule-based structure tasks. The majority could also use optimal strategies whereas the minority could use similarity strategies in 2 dimensions discontinuous rule-based structure tasks. The results in research 1 suggested that the relevant dimension number was an important influence factor to category learning.Discontinuous category stimuli distribution effects on two kinds of category learning were investigated in research 2. The discontinuous category stimuli distribution was hypothesized to make the categorization rule more complexity and impact the two kinds of category learning. The rule-based and information-integration continuous structure were compared with discontinuous structure respectively. The results showed that the discontinuous category stimuli distribution impacted not only information-integration but also rule-based category learning. The results about strategies showed that more subjects could use optimal rule strategies in simpler continuous structure tasks and more subjects could use suboptimal rule and similarity strategies in more difficulty discontinuous structure tasks.Category number effects on two kinds of category learning were investigated on the based of equalizing relevant dimensions in research 3. The results showed that category number impacted not only information-integration but also rule-based category learning. The results about strategies showed that more subjects could use optimal rule strategies in rule-based structure tasks, wherever 2-categories or 4-categories, and more subjects could use extreme value strategies in information-integration structure tasks, wherever 2-categories or 4-categories.Converging with the three researches results showed that the difficulties in classification tasks influenced not only response accuracy but also response strategies. With adding the difficulties in classification tasks, the rule was more complex, and the accuracy rates would be gone down, and the response strategies would be different. More subjects could use optimal rule strategies in simpler structure tasks and more subjects could use suboptimal rule and similarity strategies in more difficulty structure tasks. These results supported"difficulty theory"but not multiple systems theory. Our studies have provided important instruction to category learning theories and practices.