We are pleased to announce that several members of the DR2 research group have proposed a joint session with STOREP (Associazione Italiana per la Storia dell’Economia Politica / Italian Society for the History of Political Economics) during the upcoming 16th STOREP conference, to be held in Siena on June 25-27th.
Maps of Science: New Problems, New Tools, New Opportunities
Chair: Enrico Pasini (Università di Torino)
Discussants: Guido Bonino, Enrico Pasini, and Paolo Tripodi (Università di Torino)
PAPERS
Intellectual and Social Similarity among Scholarly Journals. An Exploratory Comparison of the Networks of Editors, Authors and Co-citations
Alberto Baccini (Università di Siena), Lucio Barabesi (Università di Siena), Yves Gingras (Université du Québec à Montréal), Mahdi Kelfaoui (Université du Québec à Montréal)
In this paper we consider the Interlocking editorship (IE), Interlocking Authorship (IA), an Co-Citation (CC) networks of journals of statistics, information and library science and economics. The degree of similarity among journals in the three networksis explored. The first intuitive question is whether these three network are similar. The basic idea is to explore whether the social proximity among journals observed in the network of the editorial boards is similar to the social/intellectual proximity observed in the IA network and in the intellectual proximity of the co-citation network.
The Structure of Scientific Disciplines. Some Notes on the Epistemology of Algorithmic Representations of Disciplines
Eugenio Petrovich (Università di Siena)
Representations of research fields are devices that are used to grasp the inner structure of disciplines. They play a crucial role in the disciplinary life since they ease the fulfillment of many epistemic and social tasks (from systematizing the intellectual content to organizing the cognitive labor of researchers, to manage appointments in the university).
In the first part of the paper, I argue that, from a general point of view, representing the inner structure of a field correspond to a problem of grouping. The task is to find how the disciplinary units (being them ideas, chairs or journals) can be grouped into relatively homogeneous clusters (the sub-disciplines or sub-areas). Until recently, the grouping problem has been approached by many disciplines mainly with a top-down strategy. A classificatory scheme (i.e., a list of sub-disciplines) was designed based on a variety of considerations, and then the units were classified according to its categories. Computer-based clustering algorithms developed in the last years allow now a new approach to this problem, of a bottom-up kind. Following this approach, we start by selecting some property of the units we want to classify, and then the algorithm groups the units according to their similarity and dissimilarity on this property. In this way, we let the units organize by themselves, so to say.
In the second part of the paper, I argue that the kind of structure that the bottom-up strategy captures should be interpreted differently from the structure the classic top-down approaches expose. Even if bottom-up representations can be superficially similar to the top-down ones (e.g., the clusters can be easily mapped on the traditional categories, warranting the validation of the bottom-up representation), I argue that they should be interpreted with a very different set of concepts. In particular, I will show that the concept of structure we should use to interpret them is closer to the notion of structure we find in sociology. Thus, the structure we reveal by algorithms should be intended as the latent organization, resulting from thousands of micro-actions, which shapes the degrees of freedom of the actors by imposing a set of constraints on them. This interpretation opens a range of new theoretical problems, to which the third part of the paper is devoted.
In the third section, by taking the example of field representations based on co-citation analysis1 – a widespread bottom-up technique of science mapping – I will focus on one these problems, namely the problem of how we should understand the relationship between the micro-context (the intellectual actions of the individuals) and the macro-structure revealed by the analysis. I will argue that the intermediate sub-structures, i.e. the sub-disciplines which show up in the map as clusters (denser regions), play a central role in this respect. By channeling the information into relatively homogeneous sets (the sub-disciplinary literatures), they allow the individual actors to cognitively dominate the information and thus to contribute to the production of new knowledge. By the same token, however, they constraint the possible intellectual actions of individuals: sub-disciplinary clusters exert a sort of “gravitational force” on the actors contributing to them. For instance, they render the action of “escaping” from the sub-discipline to bridge the gap with another one a costly enterprise. I will argue that they play a structuring role: they both allow and constraint the intellectual action (production of new knowledge).
If this interpretation is correct, then, bottom-up representations should be intended as revealing the latent forces (i.e., the structure) which doubly enable and constraint the intellectual action of researchers within disciplines. Understanding how these forces work and how they interact with individual actors at the micro-level is an important task for anyone who is interested in a theory of discipline dynamic based on the bottom-up representations produced by algorithms.
The Problem of Knowledge Classification in Economics: The Potentialities of Automated Information Retrieval Systems
Angela Ambrosino*, Mario Cedrini*, John B. Davis**, Stefano Fiori*, Marco Guerzoni*,*** and Massimiliano Nuccio****
* Università di Torino, Italy
** Marquette University and University of Amsterdam
*** ICRIOS, Bocconi University, Milan
**** University of Birmingham
The paper is a first, preliminary attempt to illustrate the potentialities of topic modeling as information retrieval system helping to reduce problems of overload information in the sciences, and economics in particular. Noting that some motives for the use of automated tools as information retrieval systems in economics have to do with the changing structure of the discipline itself, we argue that the standard classification system in economics developed over a hundred years ago by the American Economics Association, the Journal of Economic Literature (JEL) codes, can easily assist in detecting the major faults of unsupervised techniques and possibly provide suggestions about how to correct them. With this aim in mind, we apply to the corpus of (some 1500) “exemplary” documents for each classification of the Journal of Economics Literature Codes indicated by the American Economics Association in the “JEL codes guide” (https://www.aeaweb.org/jel/guide/jel.php) the topic-modeling technique known as Latent Dirichlet Allocation (LDA), which serves to discover the hidden (latent) thematic structure in large archives of documents, by detecting probabilistic regularities, that is trends in language text and recurring themes in the form of co-occurring words. The ambition is to propose and interpret measures of (dis)similarity between JEL codes and the LDA topics resulting from the analysis.
