The perceived decline of “disruptive” science and technology

Abstract

The discussion about science's purported loss of disruptive innovation must not distract from the underlying problems that prevent science from deploying its full potential.

A recent research article in Nature (Park et al, ²⁰²³) puts forward the hypothesis that patents and scientific publications have become less disruptive over time, which, the authors assert, would be a major concern for science and technology. This conclusion is based on the analysis of 45 million papers and 3.9 million patents with the “CD index” (Funk & Owen‐Smith, 2017), a metric which the authors use to characterize the consolidating vs. disruptive nature of a scientific paper or patent. In brief, Park et al (2023) argue that if a paper or patent is consolidating, the knowledge upon which it is built remains relevant, and subsequent work is more likely to cite both the paper and the previous publications supporting it. By contrast, a disruptive paper or patent would ultimately take citation precedence over the work it was built on. The authors then tracked citations for 5 years after each papers' publication and found a steady decline in this CD₅ index over the time period studied for both publications (1945–2010) and patents (1980–2010). This is coupled with other factors, such as an analysis of word diversity over time, which are presented in turn as evidence for an overall decline in the disruptiveness of science and technology.

Not surprisingly, this study has received substantial attention on both conventional and social media, given that many scientists do care about the quality of research, and their ability to engage in innovative research. The issue is also important to policymakers, funders, and other decision‐makers, as evidenced by the prominent positioning of impact evaluations both in grant proposals and performance evaluations.

A lot of the discussion on social media has focused on concerns that are familiar to many an academic: the ever‐expanding administrative burdens, the paucity of funding and hypercompetition, the lack of jobs which pushes the brightest minds out of academia or research, all of which distract from doing cutting‐edge or disruptive science. As an aside, it is important to note that a substantial portion of said cutting‐edge research is being done out of academia, in research institutions and industry. Nonetheless, these are of course all legitimate concerns, and these factors will indeed slowdown innovation, as researchers are being spread increasingly thin.

In conventional media, headlines abound from the more sedate “The decline of disruptive science” (https://www.insidehighered.com/news/2023/01/17/study-disruptive-science-decline) and “A decline in scientific innovation?” (https://www.science.org/content/blog-post/decline-scientific-innovation) to the more hyperbolic “Science is losing its ability to disrupt” (https://www.ft.com/content/c8bfd3da-bf9d-4f9b-ab98-e9677f109e6d), “With fewer disruptive studies, is science becoming an echo chamber?” (https://www.advancedsciencenews.com/with-fewer-disruptive-studies-is-science-becoming-an-echo-chamber/), to even alarmist subtexts like “So much for flying cars. The question is what to do about the slowdown in scientific discovery.” (https://www.grid.news/story/science/2023/01/17/groundbreaking-patents-and-scientific-discoveries-are-happening-less-and-less-often-heres-why). In its coverage of the article, Nature News quotes experts raising concern about how scientific innovation is slowing down (https://www.nature.com/articles/d41586-022-04577-5). Inside Higher Ed similarly highlights experts saying that this is “a wakeup call to institutions and funding agencies” (https://www.insidehighered.com/news/2023/01/17/study-disruptive-science-decline). Notwithstanding the media attention in the wake of the article, concerns about scientific progress slowing down have already been raised before, for instance by Cowen & Southwood (2019).

A major concern for me is the very narrow definition of innovation. Park et al (2023) essentially distinguish only between disruptive and consolidating studies. However, as Greg Satell pointed out (https://hbr.org/2017/06/the-4-types-of-innovation-and-the-problems-they-solve; Satell,²⁰¹⁷), innovation is essentially about solving problems, and there are as many ways to innovate as there are problems to solve. He distills this down to four types of innovation: breakthrough innovation, sustaining innovation, basic research, and disruptive innovation. Satell further outlines how each of these plays an important role in problem‐solving. One needs the solution that best fits the current problem; locking yourself into a single approach locks yourself out of alternative, potentially more efficient innovation strategies.

Secondly, even if one does want to focus on disruption above other forms of innovation, disruptive innovation in science and technology increasingly happens through the work of communities, not by individual researchers and papers. Trying to track disruptive science at the level of individual papers ignores that science is a team sport. Recognizing that we are moving away from the successes of the lone scientist or paper is important for how we evaluate and value science moving forward.

Furthermore, is there really a lack of innovation in science? Just look at a few examples of scientific breakthroughs that have “changed the world” in recent years. One was the great success of mRNA technology for vaccine development and delivery—even if significant progress in this area predates the COVID‐19 pandemic (Pardi et al, ²⁰¹⁸). Another example is the tremendous success of machine‐learning approaches such as AlphaFold for protein structure prediction, which have essentially revolutionized the field. Neither is new technology and would not fit into the narrow view of disruptive innovation—rather being consolidating studies under the same view—but both have “changed the world” literally and globally.

The Nobel Prize‐winning CRISPR technology has revolutionized molecular biology and biotechnology, but it stands on the shoulders of decades of research in bacterial genetics and ecology. Finally, a more personal example is the work of my late friend and colleague Dan Tawfik (Elias et al, ²⁰²¹; Jackson et al, ²⁰²²): over the course of his research career, nearly everything he did reset the field and changed how we think about science. Yet under the metric used by Park et al (2023), his research would not be considered disruptive. Despite this, as put so eloquently by Kosonocky & Ellington (2023), “We not only stand on the shoulders of this giant but will for many years to come stand in his shadow.” Are all of these examples less important than papers with a high CD₅ index in terms of changing the course of research in a particular field?

Based on the definition used by Park et al (2023), there is an apparent decline in disruptive individual papers, although the authors themselves acknowledge that they are using a novel metric that might not be ideal for quantifying what they are measuring. This does not mean, however, that scientific innovation has stopped. My own fields of computational biology and biophysics are moving forward at warp speed, the researchers I train today are learning techniques and using approaches and performing simulations that would have been science fiction 10, not to mention 20 years ago when I was a PhD student.

As highlighted by the comments on social media, there are serious reasons to be concerned, in particular in academia, about how little time researchers actually have to perform cutting‐edge research and how this impacts innovation. However, innovation across the board is alive and well, and it is too early to plan its funeral. Purporting a loss of innovative potential in this context is unhelpful, as it distracts from the very real problems faced by scientists who are held back from pushing forward the boundaries of their disciplines. Focusing on resolving those, while valuing the different types of innovation that advance the human condition, will be of more value for all of us.

Disclosure and competing interests statement

The authors declare that they have no conflict of interest.

EMBO reports (2023) 24: e56876