@article{oai:nagoya.repo.nii.ac.jp:00002146,
 author = {伊藤, 康児 and ITOH, Kohji},
 journal = {名古屋大學教育學部紀要. 教育心理学科},
 month = {Dec},
 note = {Through 1970's, Rosch and her co-workers have studied natural language and perceptual categories. The purpose of this paper is to review a series of their studies with respect to the following four parts. Part I Specification of categories; This paper refers only to substantial categories, especially to categories for the ostensive concrete objects. Attempts are made to specify the nature of category and taxonomy, and define terms used. Part II A summary of findings by Rosch and her co-workers focusing on the following four topics. 1) Prototypes; Rosch and her co-workers argued that many natural categories possess and internal structure, that is, categories are represented by prototypes or best examples and category members are defined as variations of a prototype. Prototypicality is operationally defind by a subject's judgment of the goodness of membership within the category. They presented empirical evidence that prototypicality of instances within a category correlates with many of the dependent variables used as measures for the psychological researches. For example, the more the instance becomes similar to the prototypical, a) the easier to learn, b) the easier to respond to a reaction-time task, and c) the more the production when subjects are asked to write category members. 2) Basic-level objects; Rosch and her co-workers argued that categories within the taxonomy of concrete objects are structured such that there exists one level of abstraction that represents psychologically the most basic category for classification. At this level of abstraction, instances share most attributes in common inside the category, and least attributes in commom outside of it. Empirical evidence has been presented that the same level of abstraction forms the basis in our taxonomy, where objects share number of attributes in common, the identical or similar motor programs are employed when they interact with the object, visual similarities in shapes of objects occur, and an average shape of two objects is recognizable as what it represents. 3) Family resemblance; Family resemblance is defined as possession of attributes which overlap those of other instances. Empirical evidence shows that the degree of family resemblance of instances to other instances within the category is significantly correlated with typicality ratings of these instances. 4) Context effects; Rosch and her co-workers argued that situation demands the occurrance of categorization. The further discussion in this paper suggests that both prototypes and basic levels are dependent on culture or tasks. Part III Implications for cultural studies. 1) Between subjects consistency of prototypes and basic levels are maintained, suggesting that these facts are considered to be an empirical specification of the function of commom sense, or of everyday utilization of the language. 2) A series of investigations on category by Rosch and her co-workers contributes to the improvement of methodological tools for the cross-cultural study. 3) Rosch and her co-workers also indicate that no developmental defferences are found in the ability to use basic-level objects among subjects ranging from the three-year old to the adulthood. Part IV Implications for the cognitive science. 1) This paper explored some additional factors to affect the typicality ratings of instances, such as salience, preference, prittiness, calrity, and imagability. 2) Prototypicality of instances is treated as a dependent variable with reflecting the cognitive processing, including typicality ratings of instances. Rosch and her co-workers argued that the prototypicality effect depend on the property of instances, especially on perceptual property in color categories and on family resemblance in natural categories. Occurrance of the prototypicality effect is further discussed for this paper in terms of property of instances, task demand, cognition, and frequency. 3) Findings of Rosch and her co-workers suggest that the pr, 国立情報学研究所で電子化したコンテンツを使用している。},
 pages = {17--32},
 title = {<原著>カテゴリーの研究 : Rosch を中心とする研究の概観},
 volume = {27},
 year = {1980}
}