THE PSYCHOLOGY OF CLASSIFICATION

The great complexities and rapidly changing pattern of modern knowledge have made finely‐detailed classification increasingly necessary. All the better known systems of classification, such as Dewey, U.D.C., Bliss, &c., have principles and notations which are too limited and inflexible to enable even the more general schedules to give an accurate picture of the subjects. Mechanized methods, from the simplest punched cards to the complex machines devised in the United States, do no more than sort coded, isolated concepts, singly or in groups, but not (as yet) in an ordered system; such sorting operations do not, therefore, provide a classification, that is, a reproducible preferred order. On the other hand, systems of notation, however ingenious, cannot by themselves produce the terms of a class or the true inter‐connexions between ideas, and a notation must therefore always be a secondary matter, a useful adjunct tailored to fit the order established by other means. The two main problems in classification are the achievement of a detailed analysis reflecting all the complex relationships inherent between concepts, and provision for stability of once classified material combined with perfect flexibility for the introduction of new material. It will be clear that the system of order required must be not just helpful (as Ranganathan urges) but an accurate, even though necessarily simplified, representation of the state of knowledge.