Interest in quantitative analyses of the history of the economic discipline has clearly soared in recent years. This owes, among other reasons, to the availability of both databases and new powerful analytical techniques to investigate the ever-growing corpus of economic articles published since the dawn of the discipline. The current epoch in economics, displaying characteristics that pose it in quite stark contrast with respect to the past monolithism of the core neoclassical approach, seems to require new analytical categories, with respect to those traditionally used by economic methodologists; categories enabling practitioners to understand the origins of the current fragmentation, as well as the possible persistence of the somewhat pluralistic environment of today’s mainstream economics, and even the chance of a post-foundational future for the discipline.

The aim of this paper is to illustrate the possibility to provide new interpretations of such issues by applying the topic-modelling technique called Latent Dirichlet Allocation (LDA; an unsupervised, scalable basic tool devised for discovering the hidden thematic structure of large digitalized archives of document) to the economics article stored in the full JSTOR database. The study here proposed allows discovering the most representative topics in the discipline at each decade since 1890 to 2014. This permits, in its turn, to identify the appearance and life cycles of various topics, by concentrating on the splitting and merging of topics themselves. It allows measuring the content of novelty introduced, in terms of distance between topics over time, as well as the average proximity of topics in a temporal window; and the degree of topic concentration per decade.

This package is a first step in the direction of a systematic account of the evolution of knowledge creation in economics. The hypothesis we specifically put to test at this stage concerns the relevance of specialization in its quality of (self-reinforcing) strategy to cope with the rising burden of knowledge in economics as a main engine of the current fragmentation and “mainstream pluralism”. But the study aims at exploring other issues, such as (for instance) the applied turn in economics in the Seventies, or the shifting relationships between economics and other social sciences (from economics imperialism to reverse imperialisms).