Once again, I’m brooding over science’s limits. I recently posted Q&As with three physicists with strong opinions on the topic--David Deutsch, Marcelo Gleiser and Martin Rees--as well as this column: “Is Science Infinite?” Then in March I attended a two-day brainstorming session--which I’ll call “The Session”--with 20 or so science-y folks over whether science is slowing down and what we can do about it.
The Session was inspired in part by research suggesting that scientific progress is stagnating. In “Are Ideas Getting Harder to Find?”, four economists claim that “a wide range of evidence from various industries, products, and firms show[s] that research effort is rising substantially while research productivity is declining sharply.” The economists are Nicholas Bloom, Charles Jones and Michael Webb of Stanford and John Van Reenen of MIT.
As an counter-intuitive example, they cite Moore’s Law, noting that the “number of researchers required today to achieve the famous doubling every two years of the density of computer chips is more than 18 times larger than the number required in the early 1970s.” The researchers found similar trends in research related to agriculture and medicine. More and more research on cancer and other illnesses has produced fewer and fewer lives saved.
These findings corroborate analyses presented by economists Robert Gordon in The Rise and Fall of American Growth and Tyler Cowen in The Great Stagnation. Bloom, Jones, Webb and Van Reenen also cite “The Burden of Knowledge and the ‘Death of the Renaissance Man’: Is Innovation Getting Harder?”, a 2009 paper by Benjamin Jones. He presents evidence that would-be innovators require more training and specialization to reach the frontier of a given field. Research teams are also getting bigger, and the number of patents per researcher has declined.
The economists are concerned primarily with what I would call applied science, the kind that fuels economic growth and increases wealth, health and living standards. Advances in medicine, transportation, agriculture, communication, manufacturing and so on. But their findings resonate with my claim in The End of Science that “pure” science—the effort simply to understand rather than manipulate nature--is bumping into limits. And in fact I was invited to The Session because an organizer had read my gloomy tract, which was recently republished. I had lots of reactions to The Session. Here are a few (and see also my follow-up to this post, “Is Science Hitting a Wall? Part 2):
*What’s the big bang worth? In a couple of respects, I find economists’ evaluations of scientific progress too stingy. First, they grant surprisingly little value to recent advances in digital technologies. I began my career 35 years ago writing on these contraptions called “typewriters” and doing research in places called “libraries.” I’m still blown away by my ability to access virtually infinite knowledge instantaneously from a smart phone or laptop. Yeah, fake news, spam, cyber-war, but still.
Second, economists give insufficient weight to what I consider science’s most valuable achievements, its insights into reality. The big bang theory and out-of-Africa hypothesis don’t boost gross national product, but they deepen our understanding of the world and of ourselves. Isn’t that, ultimately, what science is about?
*When science regresses. In other respects, economists’ evaluations of scientific progress are too generous. Organizers of The Session asked participants if science is slowing down. We could answer Yes, No and Other. I chose “Other,” because certain fields are arguably regressing. Psychiatrists have harmed many patients by over-prescribing medications. The cancer industry, similarly, is overdiagnosing and overtreating Americans.
Some scientists at The Session scoffed at the idea of a scientific slowdown. Biologists, pointing to CRISPR, optogenetics and other advances, were adamant that the pace of discovery is, if anything, accelerating. My response: Yes, fields like genetics and neuroscience are indeed churning out findings, but to what end? Gene therapy has been an enormous disappointment, and treatments for mental illness remain appallingly primitive.
In the realm of pure science, many physicists remain stubbornly committed to strings and multiverses, things too small and large ever to be observed. Theories of consciousness have also gotten wackier lately. Prominent experts are espousing panpsychism, which holds that consciousness might be a property of many kinds of matter, not just brains. As with strings and multiverses, panpsychism cannot be experimentally confirmed. Another sign that science is running out of gas is the sharp increase in average ages of winners of Nobel Prizes in science, and especially physics.
*The replication crisis isn’t a reporting artifact. Then there is the replication crisis, the finding by statistician John Ioannidis and others that many peer-reviewed claims cannot be replicated. Optimists insist that the scientific literature is no more flawed than it used to be. We’re just giving the flaws more attention now, which is a positive trend. At The Session, an expert on the replication crisis rejected this upbeat perspective. He said that science has become less reliable because competition among researchers for publications, grants, tenure and other rewards has become more intense.
*The Whitey-on-the-Moon Problem. Several people at The Session explained the slowdown of science with the “low-hanging fruit” metaphor. Scientists have solved relatively easy problems and moved on to harder ones, like the mind-body problem and unification of physics. Here’s a tough question: At what point, if ever, do we decide that some fruit are unreachable? Or, as an economist might put it: How much are we willing to spend on problems that have proved intractable so far? When, if ever, should we cut our losses and quit?
It depends on the problem, of course. We won’t stop investing in research on schizophrenia or cancer until they are eradicated, which might not happen for a long time, if ever. Public and private sources will surely keep supporting research that has a shot at improving our health or extending our lives. But what about investigations into the fundamental nature of matter, the origin and structure of the universe, the possible existence of extraterrestrial life, problems with no practical payoff?
To put it more bluntly: How much are “pure” discoveries like the big bang or out-of-Africa hypothesis worth? I’d like to say they are priceless, but that answer won’t suffice when we’re talking about government funding. Should we spend billions of tax dollars on a next-generation particle accelerator, gravitational-wave detector or manned mission to Mars when millions of people lack decent health care, housing and education?
I call this the Whitey-on-the-Moon Problem in honor of rap pioneer Gil Scott-Heron. In his 1970 song “Whitey on the Moon” Scott-Heron says, “A rat done bit my sister Nell/(with Whitey on the Moon)./…The man just upped my rent last night/('cause Whitey's on the moon)./No hot water, no toilets, no lights/(but Whitey's on the moon).”
*Too much innovation veneration! One driver of the replication crisis is our culture’s growing obsession with “innovation.” As technology historians Lee Vinsel and Andrew Russell state in their influential Aeon essay Hail the Maintainers: “Entire societies have come to talk about innovation as if it were an inherently desirable value, like love, fraternity, courage, beauty, dignity, or responsibility. Innovation-speak worships at the altar of change, but it rarely asks who benefits, to what end?”
Good question. Innovation veneration, Vinsel and Russell point out, has widened the gap between haves and have-nots and led to neglect of the maintenance needed to keep things running smoothly. It has driven up the costs of American medicine without appreciably improving our health. Innovation in military technologies, including drones and cyber-weapons, arguably imperils rather than enhancing our security. Researchers’ desperation to publish novel results also surely contributes to the replication crisis--and to hype escalation.
*“Metascience” versus “Limitology.” Participants of The Session proposed various names for the scientific study of science, such as “metascience.” I nominate “limitology,” which I coined in The End of Science to describe the study of the limits of knowledge. Some science enthusiasts bridle at talk of limits, and I understand why. Limitology could give ammunition to anti-science forces in our culture, and speculation that some problems are too hard to solve might become a self-fulfilling prophecy.
But some of science’s greatest insights--quantum mechanics, relativity and Godel’s theorem come to mind—impose limits on our knowledge. And since Darwin we have known that we were designed to reproduce, not to understand neural coding or quantum field theory.
*Boosterism versus realism. I once asked novelist/philosopher Rebecca Goldstein if optimism is a requirement for intellectuals. No, she replied, realism is a requirement. Some folks at The Session had optimistic, can-do, even booster-ish responses to science’s diminishing returns. They cited projects, such as the Center for Open Science and Stanford’s Meta-Research Innovation Center, aimed at making science more transparent, reliable and efficient.
I applaud these efforts too. But if limitology is to be a serious intellectual enterprise, and not an exercise in marketing, it must be based on realism, not boosterism. It must question the value of research and “innovation” and weigh the costs and benefits of different scientific enterprises. It must consider the possibility that some mysteries might be unsolvable. Ignoring science’s limits is unscientific.
Readers also sent me these comments:
Andrew Odlyzko, mathematician: I don't think science is hitting a wall. Rather, as I argued almost a quarter century ago, in the piece "The decline of unfettered research," my thesis is that as the low-hanging fruit are getting rarer, we have to devote more resources to research, and much more of it is of the incremental and predictable variety, solving problems that are selected by various policy makers. Therefore, this research is moving away from the "brilliant single investigator" mode towards a large and directed enterprise, where individuals have less freedom, and are striving to contribute to big projects.
Bob Frankston, computer scientist: Beware of economists carrying old measures. Too bad there is an attitude that research should produce a measurable result rather than appreciating that failure is the norm and the few success make up for it. The Moore’s law question is what inspired me to write “Beyond Limits” 20 years ago. The problem is typically in the measures as old veins get tapped out and we need new ways of thinking, new paradigms, for fresh measures.