Ambiguity and scientometric measurement: A dissenting view

Abe Bookstein has long been a persuasive advocate of the central role of the classical Lotka-Bradford-Zipf "laws" in bibliometrics and, subsequently, scientometrics and informetrics. In a series of often-quoted papers (Bookstein, 1977, 1990a, 1990b, 1997), he has sought to demonstrate that "Lotka-type" laws have a unique resilience to various forms of reporting, which leads inevitably and naturally to their observance in empirical informetric data collected under a wide variety of circumstances. A general statement of his position was featured in the recent JASIST Special Topic Issue on Information Science at the Millennium (Bookstein, 2001). We shall argue that there are grounds to dispute some of the logic, the mathematics, and the reality of the development. The contention is on the one hand that Bookstein's development lacks a rigorous mathematical basis, and on the other, that, in general, informetric processes are adequately described within a standard probabilistic framework with stochastic modelling offering the more productive approach.