Abstract
Two kinds of categorization are distinguished - rational and natural. Quantitative aspects of natural categorization are explored by means of a distribution first derived by Laplace but which is here slightly modified to make it more adaptable. This basic distribution, with an inverse square law probability density function, can be simply related to the empirical distri butions of Lotka, Zipf, Bradford and others. Some implications of this unified empirical law of natural categorization are discussed.

This publication has 8 references indexed in Scilit: