Science is often presented as a universal endeavor, guided by curiosity and logic rather than politics. But a sweeping new quantitative analysis of roughly 80 million publications from 1970 to 2023 tells a different story, one in which geopolitical shocks, national priorities, and deep structural inequalities between the Global North and South shape what gets studied, by whom, and for whom.
The paper, posted on arXiv by researchers at CNRS (Paris), Universite Paris Cite, and the University of Montpellier, uses the OpenAlex database to track the evolution of national research profiles across five decades. The authors, Irina Vorobeva, Maxime Lenormand, Germana Berlantini, and Floriana Gargiulo, apply tools from information theory and network science to measure how countries’ scientific fingerprints have changed over time.
Tipping points
The analysis identifies three major exogenous shocks that synchronously reconfigured global research priorities.
The Chernobyl disaster (1986) produced a sharp but geographically narrow shift, concentrating research attention on nuclear physics, toxicology, and hematology across the Soviet Union, Western Europe, and Japan. The effect was relatively brief, lasting roughly two years.
September 11, 2001, triggered a broader and more durable reorganization. Research in electrical engineering, computer networks, signal processing, and artificial intelligence surged across roughly 40 nations, a shift that, the authors note, effectively reversed the “AI winter” of the 1990s by injecting sustained defense-related funding. The effect was concentrated in Western and US-allied countries; China and Russia did not participate in this reorientation.
COVID-19 produced the most geographically diffuse shock. Research priorities shifted globally toward infectious disease modeling, environmental chemistry (for aerosol transmission studies), and even architecture (for ventilation and building design). Notably, the United States was absent from the list of countries showing the largest pandemic-driven shifts, a finding the authors attribute to the US’s already broad and diversified research portfolio, which made its relative reorientation smaller than that of more specialized countries.
The paper uses a tree-Wasserstein distance metric, a mathematical tool borrowed from optimal transport theory, to detect structural reorganization that persists longer than surface-level changes. This revealed that the COVID-19 reconfiguration, though the most recent, has already driven changes in national research profiles that exceed those of 9/11 in depth and geographic breadth.
National fingerprints
Beyond the shock-driven shifts, the analysis reveals persistent “national fingerprints”, the unique relative distribution of each country’s scientific output across disciplines. The researchers measure the distance between these fingerprints using Jensen-Shannon divergence, tracking which countries converge toward a shared global agenda and which maintain distinctive local priorities.
The United States shows a remarkably stable research profile over five decades, apart from the 9/11 shift, a reflection of its role as the dominant scientific superpower with a broad, diversified portfolio. Russia shows abrupt shifts coinciding with the 1991 dissolution of the Soviet Union and again in 2014 and 2021. China’s profile shifts periodically, aligning with its five-year planning cycles. Brazil underwent a significant reorientation in 2007, coinciding with President Lula’s PACTI science policy that expanded research funding.
Unequal globalization
The most striking finding concerns the persistent gap between Global North and South. While the domestic share of publications has declined globally, reflecting the rise of international collaboration, the pattern is deeply unequal.
For most countries in the Global South, international collaborations dominate their publication output, and the content of those international papers diverges sharply from the content of their domestic research. Southern intellectual resources, the authors argue, are mobilized for externally defined agendas. They invoke the concept of “scientific extraversion”, a term from Beninese philosopher Paulin Hountondji, to describe how Southern researchers’ international output is shaped more by the priorities of their Northern partners than by local needs or traditions.
A small number of countries show signs of what the authors call “scientific emancipation”, Indonesia, Iran, and Morocco have all increased their domestic publication share while reducing the divergence between their domestic and international research profiles. But these are exceptions.
A polycentric future
The analysis also tracks which countries’ research profiles are emulated by others, what the authors call “aspiration networks.” Using PageRank centrality (the same algorithm Google uses to rank web pages), they find that China’s centrality has risen dramatically since 2000, while Japan, Russia, and major European powers have declined. Brazil, India, and Indonesia are rising. The former unipolar Western core is giving way to a polycentric structure.
Gravity-model analysis confirms the trend: in 1983, the largest research producers had the most similar profiles, a tight, Western-dominated core. By 2023, the largest producers were more distant from each other, suggesting increasing global scientific polarization.
The paper is available on arXiv under a CC BY 4.0 license.
Disclosure: Based on an arXiv preprint that has not undergone peer review.
Sources
[1] Vorobeva, I., Lenormand, M., Berlantini, G., & Gargiulo, F. “The geopolitics of knowledge: tipping points, national fingerprints, and the unequal globalization of science.” arXiv:2607.08512 (2026). https://arxiv.org/abs/2607.08512