“What science cannot discover,” author Bertrand Russell told us, “mankind cannot know.”

Russell’s claim, which is a philosophical and a not-scientific assertion (thus, it refutes itself), reflects the view commonly known as “scientism.” This is the premise that science is the best, perhaps even the only, means of getting truth. Or as evolutionary biologist Austin L. Hughes wrote in an article called “The Folly of Scientism,” it’s the idea, held by many scientists, “that natural science does or soon will constitute the entire domain of truth.”

And it’s not just scientists, either. A year ago I spoke at a meeting of church workers. When I asked, “What is the most certain mode of knowledge we have?” a man blurted out, “Science.”

It’s easy to understand why people would believe such an illusion (though one would think a church worker would know better). After all, think about the vast domains of knowledge that science has opened up to us. From the structure of the atom to the structure of the universe, it is giving us glimpses of reality (or at least our experience of reality) that we haven’t received from anywhere else. Look at all the technology, surely the lushest fruit plucked from the tree of scientific knowledge.

Yet science is a much more subjective endeavor than most people realize. What about, for instance, the many scientific theories or even the assumptions needed to do the science—all of which were once deemed certain—that have been cast by the wayside? In physics alone many theories and assumptions that were accepted at the beginning of the twentieth century as unquestionable never made it to the twenty-first.

Before the twentieth century, science taught that the universe was eternal. That belief is gone. Before the twentieth century, science taught that the universe was static. Gone. Before the twentieth century, science taught that time and space were absolute. Gone, swept away by the mind of Albert Einstein.

Before the twentieth century, science taught that the universe was deterministic. This is the idea that the future course of physical events, even down to the tiniest level, were fixed by past causes. Determinism was an essential assumption of science, yet within a few decades of the twentieth century it was gone. Another fundamental assumption was the continuous nature of physical process, that all changes in the physical world, however great, could be analyzed in terms of a continuous sequence of ever-smaller changes. Gone. Before the twentieth century, it was just assumed that the physical state of any composite system was nothing more than the sum of its parts, and that any system, no matter how complex, could be fully explained by a study of those parts. That view, so fundamental that it was never recognized as such until challenged by quantum mechanics, never made it to 2000 either.

No wonder Alfred North Whitehead (1861-1947) could write: “Fifty-seven years ago it was when I was a young man in the University of Cambridge. I was taught science and mathematics by brilliant men and I did well in them; since the turn of the century I have lived to see every one of the basic assumptions of both set aside. . . . And yet, in the face of that, the discoverers of the new hypotheses in science are declaring, ‘Now at last, we have certitude.’ ”

Basic assumptions turned aside? This leads to an insuperable weakness of science—and that is the certainty (or uncertainty) of its premises. Science, like all epistemological endeavors, must proceed from assumptions. But how do you validate what you must assume? This becomes especially problematic because, at the most fundamental level, scientific assumptions are, essentially, philosophical. And when these get overturned, a lot of science once deemed certain gets overturned with them.

“The theories we currently hold to be true,” said Steven Goldman, a philosopher of science, “are as likely to be falsified in the next 100 years as the theories that we look back on as having been falsified in the last 100 years.”

Sure, science teaches us a lot about the world (or at least how we experience the world). But the idea that it’s the ultimate arbiter of truth is a philosophical, not a scientific, claim—and a flawed one at that.