2023 had barely begun when scientists got some jolting news. On Jan. 4, a paper appeared in Nature claiming that disruptive scientific findings have been waning since 1945. An accompanying graph showed all fields on a steep downhill slide.
Scientists took this as an affront. The New York Times interpreted the study to mean that scientists aren’t producing as many “real breakthroughs” or “intellectual leaps” or “pioneering discoveries.” That seems paradoxical when each year brings a new crop of exciting findings.
In the 12 months following that paper, scientists have listened to the close encounters between supermassive black holes, demonstrated the power of new weight loss drugs and brought to market life-changing gene therapies for sickle cell disease.
What the authors of the January paper measured was a changing pattern in the way papers were cited. They created an index of disruptiveness that measured how much a finding marked a break with the past. A more disruptive paper would be cited by many future papers while previous papers in the same area would be cited less — presumably because they were rendered obsolete. This pattern, they found, has been on a decadeslong decline.
One of the authors, Russell Funk of the Carlson School of Management at the University of Minnesota, said they wanted to measure how new findings shifted attention away from old ways of doing things: “Science definitely benefits from a cumulative work and studies that come along and refine our existing ideas. But it also benefits from being shaken up every now and then.”
Scientists I talked to said researchers cite papers for many reasons — including as a way to ingratiate themselves with colleagues, mentors or advisers. Papers on techniques get a disproportionate number of citations, as do review articles because they’re easier to cite than going back to the original discoveries.
Biologist Gregory Petsko of Harvard Medical School said a better way to think about important science would be to consider some findings transformative — opening new avenues without closing off the old — though he agreed that the funding agents could get more disruption by taking more chances on long-shots.
The world does need more long-shot research, said George Church, professor of genetics at the Harvard-MIT Program in Health Sciences and Technology. “In fact, you should fail a million times a day, which is in contrast to the NASA motto, which is failure is not an option,” he said.
He strongly disagrees with the notion that scientific progress is declining. He co-wrote a rebuttal to the disruption paper for STAT, in which he argued disruption as measured by the study doesn’t reflect what we should really want from science, which is knowledge that can help us live longer, better, healthier lives.
Brian Uzzi, a professor at the Kellogg School of Management at Northwestern University, had another explanation for the changing pattern in scientific discovery. He said one thing that’s changed steadily since 1945 is our cumulative knowledge. “Every year, more papers get published than the year before,” he said. Now there are more than a million a year. That means, by necessity, students are trained more narrowly and are equipped to see smaller pieces of big problems.
“Revolutions are less likely to come down to individual papers; today they just happen on a more gradual scale as different researchers take on different pieces in a divide-and-conquer fashion.”
With the complexity of many of today’s particle accelerators and space telescopes it’s tempting to consider that most of the easy problems have been solved — an argument that’s mocked whenever it’s made.
But even if it’s true that the foundations of biology and physics are never going to be toppled, there’s plenty of science yet to be done that most of us would consider profound — especially in the applied sciences.
From curing disease to reducing global warming, there’s no shortage of hard scientific problems crying out for solutions.
F.D. Flam is a Bloomberg Opinion columnist covering science.
