Journal Article
Machine Learning

From translations to problematic networks: An introduction to co-word analysis

Michel Callon(Institut Méditerranéen des Sciences de l'Information et de la Communication), Jean-Pierre Courtial, William A. Turner, Serge Bauin
March 1, 1983Social Science Information1,894 citations
PDFDOI

1.9k

Citations

44

Influential Citations

Social Science Information

Venue

1983

Year

Abstract

This article is the story of a summer shock, my encounter with the manuscript of a book. And what happened little by little. A fragmented and incomplete narrative. He wanted to see things “irreducible and festive”! That's Bruno Latour's youthful, resonant cry.And so it was, my electroshock when I read “Irréductions”, the second part of “Microbes, War and Peace” « The pasteurization France (1984), on a mediterranean beach in July 1981, under a friendly blue sun, and a breeze for playful sails. In the previous decade, I had already been shaken, oh so much! by a thunderclap in an August-lit sky, by the publication of Anti-Œdipe: 1972. The shock was renewed nine years later with the publication of Mille Plateaux. What a decade and what a world, beginning with the Soviet army's entry into the war in Afghanistan! A living crack in « Communism in action », leading to its collapse. This article also attempts to express the gradual, hesitant emergence of doubt about the effects of insomniac reason. And of the development of “hypercontrol” societies.

Analysis

Why This Paper Matters

This paper, published in 1983, is a seminal contribution to the field of scientometrics and the sociology of science. It introduces co-word analysis, a method that uses the co-occurrence of keywords in scientific publications to map the structure and dynamics of research fields. The paper is significant because it provides a quantitative, network-based approach to understanding how scientific knowledge is organized and evolves over time, moving beyond qualitative case studies.

The work is deeply rooted in actor-network theory (ANT), co-developed by Bruno Latour and Michel Callon. It conceptualizes scientific activity as a process of translation and network building, where actors (researchers, institutions, concepts) are linked through shared problems and solutions. Co-word analysis operationalizes this theory by treating keywords as traces of these translations, allowing researchers to visualize the 'problematization' of a field.

Technical Contributions

  • Co-word analysis framework: Formalizes the extraction of keyword co-occurrence data from scientific literature and the construction of co-word networks.
  • Network measures: Introduces indices such as inclusion index and proximity index to quantify the strength of association between keywords.
  • Dynamic mapping: Shows how co-word networks can be compared over time to identify emerging research fronts, stable cores, and declining areas.
  • Link to ANT: Provides a methodological bridge between the qualitative concepts of actor-network theory and quantitative network analysis.

Results

The paper does not present quantitative results in the traditional sense (e.g., accuracy metrics). Instead, it demonstrates the method through illustrative examples, such as mapping the field of polymer chemistry. These examples show that co-word networks can reveal the central role of certain concepts (e.g., 'catalysis') and the clustering of research around specific problematizations. The method's validity is argued through its ability to reproduce known historical developments in science.

Significance

Co-word analysis has had a lasting impact on the field of scientometrics and beyond. It is a precursor to modern science mapping techniques, including co-citation analysis and bibliographic coupling. The method has been widely used for research evaluation, technology foresight, and innovation policy. The paper's integration of quantitative methods with sociological theory has influenced subsequent work in science and technology studies (STS). Today, co-word analysis remains a standard tool for analyzing large-scale scientific corpora, and its principles underpin many contemporary bibliometric and altmetric approaches.

co-word-analysisscientometricsbibliometricsnetwork-analysisscience-mappingactor-network-theory

Related Papers