Is modern science evolving in the wrong direction?

Publishing scientific journals online entails not only casting papers in PDF, HTML, and other online formats. The process also turns a paper’s title, authors, author affiliations, subject areas, and other information into index entries in searchable databases. References and citations also become searchable.

Thanks to that trove of data, scholars can investigate trends, correlations, and connections that would be harder to spot in print publications. In 2005 Sidney Redner, who studies diffusion processes, complex networks, and other manifestations of many-body interactions at Boston University, crunched through 110 years of citation data from the Physical Review family of journals.

Published in Physics Today‘s the June 2005 issue, Redner’s analysis revealed, among other things, that in the period covered by his study, June 1893–July 2003, papers received 8.8 citations on average. Nearly 70% of papers received fewer than 10. My favorite finding: A selection of six landmark papers that each had around 1000 citations had startlingly different citation histories. Whereas citations to John Bardeen, Leon Cooper and Robert Schrieffer’s “Microscopic theory of superconductivity” of 1958 more or less tracked the rise and fall from fashion of type-1 superconductors, citations to Walter Kohn and Lou Sham’s exposition of density functional theory has been rising monotonically ever since its publication in 1965.

The latest example of bibliometric analysis to attract my attention is a preprint by Klaus Jaffe of Simon Bolivar University in Caracas, Venezuela. Under the title “Is modern science evolving in the wrong direction?,” Jaffe looks at publishing trends in several disciplines during the past 12 years.

In his introduction, Jaffe approvingly cites a recent report by the Royal Society of London that touts several benefits of international collaboration in science. Those benefits, which include mutual economic growth, are not universally recognized, it seems. Jaffe found, for example, that whereas papers by scientists in Western Europe had a higher fraction of international coauthors in 2011 (36%) than in 1999 (25%), papers by scientists in Asia had the same, lower faction in both years (15%).

The results for physics were mixed. Of all scientists in all regions, physicists in Western Europe were the most likely to have international coauthors. In 2011 52% of their papers had at least one coauthor from another country. In Asia, the fraction was lower, 24%, but it significantly exceeded the regional average of 15%. On the other hand, perhaps because physicists already collaborate across borders at high rates, physics exhibited one of the smallest increases in international collaboration from 1999 to 2011.

Another trend that Jaffe found is that international collaboration was strongest and grew faster in basic sciences than in applied sciences. That difference makes sense to me. One common mode of collaboration—perhaps the most common—involves experimenters working with theorists. Such partnerships are more likely in basic science than in applied science because the closer a project is to an application, the more likely the relevant theory has been established.

Because he used citation data, Jaffe could not readily identify what might be called local international collaborations, which occur when scientists from different countries work together in the same country. That kind of collaboration is so common that we perhaps overlook it. But collaboration between people of different national backgrounds is fruitful regardless of where the people live.

Science transcends national boundaries because its subject, the workings of nature, disregards borders. Moreover, the most common qualification to be a professional scientist, the PhD, is more or less internationally recognized. Given those advantages, it would be a pity if the trends spotted by Jaffe persisted and strengthened.