Natural categories: Well defined or fuzzy sets?

MICHAEL E. McCLOSKEY 0 1 SAM GLUCKSBERG 0 1 0 This work was done while the first author held a National Science Foundation predoctoral fellowship. The research was supported by Public Health Service Research Grant MH 23401 , S. Glucksberg, principal investigator. We thank Nancy McCloskey for her help in data analysis. Requests for reprints should be sent to Michael McOoskey, Psychology Department, Princeton University , Princeton, New Jersey 08540 1 Princeton University , Princeton, New Jersey 08540 Thirty college students made category membership decisions for each of 540 candidate exemplar-category name pairs (e.g., apple-fruit) in each of two separate sessions. For highly typical category members (e.g., chair for the furniture category), and for items unrelated to a category (e.g., cucumber-furniture), subjects agreed with each other and were consistent in their decisions. However, for intermediate-typicality items (e.g., bookends-furniture), subjects disagreed with each other and were frequently inconsistent from one session to the next. These data suggest that natural categories are fuzzy sets, with no clear boundaries separating category members from nonmembers. - METHOD Subjects Sixty-four men and women undergraduates at Princeton University served as paid volunteers. Thirty subjects served in the primary category membership decision group, 24 subjects provided typicality ratings for the stimulus pairs, and 10 subjects made exemplar-category partial overlap judgments. RESULTS Table 1 Distribution of Candidate-Exemplar/Category-Name Pairs Across Typicality Levels Intermediate Typicality Level Corresponding Typicality Ratings 9.00-10.00 8.00- 8.99 7.00- 7.99 6.00- 6.99 5.00- 5.99 4.00- 4.99 3.00- 3.99 2.00- 2.99 1.00- 1.99 Number of Pairs (W/) .8 z(/) Oz ,6 - 0 I-a. Oa:(w/) .4 7 6 5 4 3 TYPICALITY LEVEL TYPICALITY LEVEL 7 6 5 4 3 TYPICALITY LEVEL *p < JJ5 Categories Per Name 20.83* 1.63 9.63* 13.33* 20.83* 24.30* 14.70* 28.03* 2.70 28.03* 20.83* 28.03* 14.70* 14.70* 9.63* 28.03* 17.63* 20.83* DISCUSSION Appendix Data for Individual Stimuli Candidate Exemplar Typi cality Typicality Carpenter's Tool Category y Clothing Category Candidate Exemplar Candidate Exemplar Typi cality Typicality Candidate Exemplar Candidate Exemplar Typicality Kitchen Utensil Category 9.42 y .02 9.38 y .02 9.25 y .03 8.58 y .02 8.54 y .05 8.25 y .20 8.21 y .13 8.13 y .25 8.04 N .47 7.38 y .33 7.33 y .13 7.21 y .20 7.13 y .25 6.92 N .47 6.54 y .45 6.46 y .30 6.42 y .37 Typicality Science Categnry y .00 Precious Stone Category Candidate Exemplar Typicality Candidate Exemplar Chemistry Biology Physics Astronomy Oceanography Medicine Anatomy Meteorology Mineralogy Psychology Engineering Metallurgy Dentistry Nursing Pharmacy Nutrition Archaeology Anthropology Geometry Agriculture Criminology Economics Geography Architecture Sociology Linguistics Politics Philosophy History Advertising Sport Category Candidate Exemplar Typi cality Vegetable Category Vehicle Category Candidate Exemplar Typi cality *MR = modal response; NR x: proportion of nonmodal responses; WI = within-subjects inconsistencies. REFERENCE NOTE J. McCloskey. M .. &: Glucksberg. S. Decision processes 11l I'eri!ring class inclusion statements: Implications for models 01 semantic memory, Manuscript submitted "for publication. IlC. Weather Phenomenon Category (...truncated)