When the news is full of vaccine-denial, climate crisis-denial, and people vowing they’ll go to their graves to defend their right to pollute their own homes with toxic fumes, it’s easy to believe that the Age of Reason has been filed away in the history books. However, when every year seems to bring faster computers, a better cell phone, more practical electric vehicles, and some new gadget that you never knew you needed before it appeared … can we really be looking at the end of the scientific age?
The whole idea seems kind of silly. It’s not as if we’re about to just turn off all the light switches and go hunting for the nearest cave (Which are rarely even listed on Zillow).
However, there’s a very serious issue at the core of this question, one that was recently taken up by an article in Nature, one of the most respected scientific journals in the world. That paper suggests that while the number of new scientific publications has never been higher, the impact of those publications is constantly declining. The average new patent application is upsetting Einstein and discovering faster-than-light travel. The latest papers are not sending the Standard Model packing with astounding new results out of the world’s largest particle colliders. Instead, we’re just sort of polishing up the edges of theories that are decades old, and the results of that polishing are, on average, increasingly inconsequential.
There are reasons to be skeptical of the paper in Nature, and scientists writing in response have been quick to point them out. First, the authors aren’t scientists in the common sense; they have doctorates in business management, not biology. Second, the data they used wasn’t polling of those in the fields, but a survey of patent filings.
This has promoted criticism along the lines that the paper is more about the lack of new ideas busting open new areas for business, than in truly assessing changes to the underlying science. However, the actual data seems compelling, and the impact could hardly be higher.
We find that papers and patents are increasingly less likely to break with the past in ways that push science and technology in new directions. This pattern holds universally across fields and is robust across multiple different citation- and text-based metrics. Subsequently, we link this decline in disruptiveness to a narrowing in the use of previous knowledge, allowing us to reconcile the patterns we observe with the ‘shoulders of giants’ view. We find that the observed declines are unlikely to be driven by changes in the quality of published science, citation practices or field-specific factors. Overall, our results suggest that slowing rates of disruption may reflect a fundamental shift in the nature of science and technology.
The bulk of the paper is related to how they determined “disruptiveness” of papers and patent filings (which is where many of those offended by the idea find traction in disputing the overall theme), but the thrust of the conclusion is this: The number of publications has increased, many of those papers are very high quality, some remain disruptive, but many only confirm the status quo. Or at best, they offer new insight that leads to little potential for either scientific or economic impact.
This has immediately caused many publications covering the story to go in one of two directions. First, there’s doom mode (if not DOOM mode), expressing concerns that the well of new ideas may be running dry and that we’re genuinely running out of new things to discover. This is sometimes followed by pondering which STEM fields will die first.
The second response usually starts with the phrase “in the late 19th century” before explaining how relativity and quantum mechanics upset the Newtonian apple cart. The last time we were convinced we knew everything, goes this line of reasoning, it turned out we really knew next to nothing.
To this second idea, the only possible response is: Boy, I wish! Nothing excites a scientist, any scientist, more than results that don’t fit the prediction. For decades now, thousands of researchers have worked diligently to poke holes in the standard model of particle physics, the limits of relativity, and the fundamental frustrations inherent in quantum. However, each time a potential deviation from the predictions of the models appears, subsequent research only seems to serve to discover flaws, not in the basic theories, but with the previous paper. We are definitely living in an age where some of the predictions made a century ago have been tested to apparent equation … and stubbornly turned upright again and again.
There are, of course, some well-known holes remaining. Wedding quantum to relativity remains elusive, even though schemes for making it work on (unvetted) paper are never in short supply. On a large, cosmological scale, making the visible universe obey our equations requires the belief that the vast majority of everything exists as invisible matter and inexplicable energy. These are giant Fudge Factors of the first water, and it seems just as likely that Dark Energy and Dark Matter will be written out of existence by some future insight into the math as it does that either will be “discovered” in a meaningful way. However, some of the most appealing theories that might offer fresh insights, often stitched together by thousands of scientists working over multiple decades, keep failing to hold up to real-world testing (i.e., supersymmetry and string theory).
There’s an important precursor to this paper that many media seem to have omitted from this discussion, and that’s the 1996 book, The End of Science, by science journalist John Horgan. Horgan is a prolific author and columnist for Scientific American, whose interview subjects included a list of scientists who might be seen as the most disruptive crew of the last three generations, from E.O. Wilson and Roger Penrose to Richard Dawkins, Stephen Jay Gould, and Stephen Hawking. Horgan has talked to them all, and hundreds of others.
Horgan’s book was a surprise bestseller (any time a science book makes the bestseller list, it’s a surprise), but there was a hard cadre of both scientists and science journalists who took umbrage at the book’s core idea: We should expect fewer, and less important, scientific discoveries as time goes on.
The reasoning behind this was simple. In the beginning, everything was available to discover. Scientists could make a discovery about the scale of the Earth with an upright stick. They could learn about the speed of sound by watching someone chop wood. However, with each passing year, as the big book of facts became more stuffed with learning, the difficulty of making fundamental new discoveries increases. In the 19th century, the electron was discovered by one guy using equipment that might have been found in a high school science lab (or the basement of a wealthy naturalist). To close out the particle zoo with the Higgs Boson took an international effort with an over $4 billion collider.
Seriously, how hard is it to believe that the easy stuff has been picked clean, and the discoveries that remain will come at an increasing cost? It seems entirely reasonable. But it was the next step that left a painful lump in many throats: What happens when the cost of a new discovery becomes so high that it simply is not achieved? Horgan saw that day if not already at hand, the certainly right around the corner. Horgan meshed this with a disquieting reminder of something that echoes the discussion between God and Job: If there are any rules by which the universe operates, there is no guarantee that we evolved apes can comprehend those rules in a meaningful way.
… given the limits constraining further research, science will be hard-pressed to make any truly profound additions to the knowledge it has already generated. Further research may yield no more great revelations or revolutions but only incremental returns.
There’s an extremely uncomfortable agreement between this conclusion from Horgan and the overview of the paper published in Nature. Or at least, in the way that paper is being presented in most of the media. However, that’s not a very accurate description of what’s actually hiding at the end of the patents and papers review.
Because while the number of revolutionary papers being published may have declined as a percentage of the total, and the impact of new research may even be somewhat fading as a whole, there is still a healthy vein of disruption in what’s being published today.
… the stability we observe in the sheer number of disruptive papers and patents suggests that science and technology do not appear to have reached the end of the ‘endless frontier.’
The implications of this go beyond just breathing a sigh of relief over finding new features in the iPhone 15. It may not be obvious, but the fundamental scientific theories that undergird our view of the universe remain open to attack. And that’s a good thing.
Hell yeah! Election season is already here, and it's already off to an amazing start with Democrats' huge flip of a critical seat in the Virginia state Senate, which kicks off this week's episode of The Downballot. Co-hosts David Nir and David Beard dissect what Aaron Rouse's victory means for November (abortion is still issue #1!) when every seat in the legislature will be on the ballot. They also discuss big goings-on in two U.S. Senate races: California, where Rep. Katie Porter just became the first Democrat to kick off a bid despite Sen. Dianne Feinstein's lack of a decision about her own future, and Michigan, which just saw veteran Democratic Sen. Debbie Stabenow announce her retirement.