Abstract
Many studies have shown that holistic feature (e.g., shape) is focused more than analytic feature (e.g., color) in visual objects recognition and noun category generalization. In the present study, we hypothesize if object segregation involves with the holistic feature dominance by investigating discovery of artificially defined "object". In experiments, participants need to segregate a target pair of elements (i.e., "object") from a distractor element in order to answer appropriate feature of the target. Target pairs are defined by conjunction of holistic and analytic features. Participants are required to learn a set of target pairs with unknown features based on feedbacks to their responses. The learning patterns suggest that object segregation is relevant to holistic feature dominance, because holistic feature was focused more in a well-separable condition when elements were sparsely distributed, meanwhile, analytic feature was focused more in a ill-separable condition when elements were densely distributed.
