The Knowledge Machine That Failed

The debate about science and its place in public policy goes way back. When President Obama pledged in 2009 to “restore science to its rightful place,” he merely begged the question: What is the rightful place of science?

In their more extreme forms, postmodernists deny the possibility of truth that is validated by repeated observations, even as they make use of scientific discoveries in their everyday lives. On this view, science has no “rightful place” because nothing can be right or wrong. The conservative critique of science is different. It sees science as a fruitful and objective pursuit, despite challenges like replicability, but denies that it can be easily translated into good public policy given uncertainties, risks, costs, moral pluralism, and unknown futures. Conservatives are not skeptical of science so much as they are of Science.

In 2020, on the eve of the COVID pandemic, New York University philosopher Michael Strevens was putting the finishing touches on his book The Knowledge Machine, a paean to modern science. The simple thesis of this work is that modern science occurred not because of social, economic, and political developments unique to the West but because of a sudden rejection of philosophy as a means to truth in favor of empirical observation alone. The rules of science that resulted were then “fixed for all time”—data, data, data—and would never again be questioned.

Read today, this thesis has more than a tad of pre-COVID about it. Indeed, the book contains some pre-COVID references that are like entering a time machine. Descartes’ “ingenious but baroque” theory of matter, Strevens wrote, was like the “jostles and shoves” aboard “certain New York City subway lines at rush hour.” By the time the book was released in October 2020, ridership on this same New York subway was just one fifth of its pre-pandemic level. Today it is stuck at 60%, as subway crime has soared and dispersed work has taken hold.

[Related: “Science: Are we getting what we’re paying for?”]

COVID not only rendered this metaphor for Cartesian matter theory “ingenious but baroque,” but it also exploded the book’s central idea of a perfected modern science. After three years of deep division on what the COVID science says and how to make use of it, describing modern science as having reached a state of perfection is baroque itself. It wasn’t only the New York subway system that broke down in 2020. The “knowledge machine” that is the object of Strevens’ adoration broke down as well.

Read one way, Strevens’ book is political propaganda designed to force-fit science into left-wing policies in areas like climate change, evolution, and public health. An Australian radio commentator made explicit what seems to be the underlying animus of Strevens’ book: “After four years of the Trump administration, I suspect that almost every scientist left in North America is aware that demagogues in politics, religion and business are conducting a lucrative war against research results.”

Yet the things that conservatives and Trumpists were saying about science had long been discussed by academics on the Left. I don’t mean the postmodernists, but the empiricists who realized that science was entering uncharted territory. As early as their 1990 book Uncertainty and Quality in Science for Policy, Funtowicz and Ravetz warned that science was rapidly reaching its limits because of its internal complexity and because its application depended on making assumptions about the future that were simply unknowable. “Ch-Ch Syndrome,” referring to the Chernobyl and Challenger disasters, signified the need for better human control over complex or unpredictable technologies.

Bjørn Lomborg’s The Skeptical Environmentalist (1998) was a critical first application of this new approach, one that was widely accepted within the policy community for highlighting issues that science did not. The application of “settled science” to the human world was running into more and more obstacles. It was contested in fundamental ways, it did not work, it created unexpected problems, or it simply wasn’t accepted. We would never again be in a “normal science” mode, to use Thomas Kuhn’s phrase for the long periods when scientists all accepted a single approach. Knowledge had reached a point where it was so complex, and affected so many different values in so many different contexts, that there could be no straight-forward consensus on what was true. “Settled science” was dead on many new and emerging issues, even if its past achievements remained uncontested. While Trump would no doubt be happy to claim the kill, it was dead for many years before.

[Related: “The Racialization of a Top Science Journal”]

The holes in Strevens’ thesis are most clear where he lets his cool façade slip by denouncing anyone who rejects the idea that incremental Darwinian modifications are sufficient to explain the complexity of life on Earth. While noting that “perhaps a third” (later reduced to “many” and then “a handful”) of scientists still believe in God, Stevens has no doubt that the knowledge machine will shortly render them naked before the truth of Darwin. He might have spent some time with his secular colleague Thomas Nagel, who in his book Mind and Cosmos calls the Darwinian theory untenable because of the impossibility of consciousness arising from incremental modifications. Others note the impossible convergence of several constants in the universe that sustain life. But that only underlies the rather quaint praise of modern science in Strevens’ book that was outdated the moment it was published. His dismissal of nineteenth-century British polymath William Whewell’s hope to integrate science and faith as “pure gobbledygook … to my philosophical ear” says more about the tin content of the latter than the empirical basis of the former.

What to do? Some of the fixes belong to science itself, and some do not. Of course, replicability, data transparency, interest disclosures, etc. all matter. But the issues go beyond research design and methods. Let’s say it is true that human-induced climate change will have major effects on the environment. So what? That tells us nothing about the effects on humans, which are presumably why we care about it, notwithstanding the perfectly reasonable case for caring about non-human life. To know climate change’s effects on humans requires making large assumptions about future states of the human world—its size, technology, wealth, organization, and values—that are well beyond the scope of ice cores from Baffin Island.

One solution is the notion of “extended peer review,” in which these multiple conflicting values are brought to bear on scientific questions. That is one way that science can be used for knowledge. The last line of Strevens’ book is, “Let us hope that knowledge saves us.” Amen to that.

Image: Adobe Stock


  • Bruce Gilley

    Bruce Gilley is a professor of political science at Portland State University and a member of the board of the National Association of Scholars. In addition to his work on academic freedom and the revival of intellectual pluralism on campus, Dr. Gilley’s research centers on comparative development and politics as well as contemporary public policy issues.

    View all posts

5 thoughts on “The Knowledge Machine That Failed

  1. There is a lot of settled science and will continue to be a lot on many relatively simply issues: pasteurized milk keeps for longer. Does anyone doubt this? Capitalism and science have made our life immeasurably better. No doubt. But, as science takes on more complex topics amidst a more complex world, the high ideals of Modern Science are less successful. Humility is needed now. If the Left was as committed to “settled science” as it says, it would be embracing nuclear power, GMO food, and limits on Round-Up and talcum powder litigation.

    1. But unpasteurized milk tastes better — and while I’ll admit to having a case of the ultra-pasteurized boxed milk stashed away for snowstorms (or zombie attacks), I throw pasteurized milk away before it expires because it doesn’t taste quite right.

      So while on one hand the science may be “settled” on pasteurized milk, if you buy it from a farmer whom you know, who keeps a clean barn and does everything “right”, the science on that is “settled” as well — and one of the big issues is do-good health officials attempting to prevent the sale/consumption of raw milk.

      So we have “science” on both sides — yes, pasteurization will kill the bad stuff in milk (which, more than shelf life, is why it was introduced) but if you keep a clean barn and don’t put the bad stuff into the milk in the first place, raw milk is also perfectly safe.

      And let us not forget that “science” *still* says that DDT is an effective insecticide. It is — it just also accumulates in the food chain… Asbestos (if kept dry) *is* an effective insulator against extreme heat, and before some of the synthetics we have now, essentially the ONLY such insulator we had. It protected people from 600 degree steam lines and firefighters (steel mill workers, etc) from extreme heat. And we didn’t have mold problems the way we do now when we put lead and mercury in paint because they kill it — except that heavy metals also aren’t too terribly good for humans, either. And irradiating meat *would* kill any/all EColi on it, but that didn’t really catch on, did it?

      My point is that even the simple settled science isn’t all that settled.

    2. Thanks to Bruce Gilley for bringing attention to Michael Strevens’ book, hadn’t heard of it, but it sounds interesting. I will give it a look.

  2. PhD stands for Doctor of Philosophy — why are we awarding that to scientists?

    Do we consider science to be philosophy or a liberal art? Or do we consider it to be a science? (On the undergrad level, we have the BS instead of the BA, and the MS instead of the MA.)

    Semantics, yes — but also maybe not.

    But the larger thing is that “settled science” does not exist and never has existed — science is the “best theory we have at the present time” with all the scientists attempting to find a better theory. Anyone remember a man named Einstein?

  3. The knowledge machine that failed? I don’t know, quantum mechanics still seems pretty good, even if nobody really understands it. (I’m not sure that quantum mechanics can be said to have come from a “knowledge machine.”) On the other hand, a lot of medical science really isn’t. But it didn’t take covid to uncover that. It was known before covid that nobody really understands cancer, nobody understands Alzheimer’s much at all. But the knowledge machine will keep plugging away, and someday I’ll bet there will be bigger progress. Covid? I think mRNA vaccines show the knowledge machine at one of its great moments. I was certainly thrilled when I got my first shot. I was very grateful for having been tripled vaxxed when I finally did get covid, which was very mild.

    In Bruce Gilley’s home state of Oregon, the covid death rate was actually less than half of the national covid death rate. Just as was the case in Canada. In some other places like Norway and Japan, it was much smaller still. It seems to me there must have been people in those places paying attention to science. At least, up until the whole world got tired of worrying about covid. Now none of those places seems to be doing better than normal.

Leave a Reply

Your email address will not be published. Required fields are marked *