Author: Michael Zerella
Category: Philosophy of Science
Word Count: 1000
“A million successful experiments cannot prove a theory correct, but one failed experiment can prove a theory wrong.”
Perhaps you’ve heard someone use this cliché to describe the scientific method as a tough-minded and unsentimental pursuit of an accurate understanding of nature. The sentiment has its roots in Karl Popper’s mid-20th-Century account of scientific investigation called “falsificationism,” so it is perhaps unsurprising that Popper’s views have been popular among many proponents of science. Unfortunately, if we are to take the cliché literally, and in the way Popper intended, the central dictum of falsificationism turns out to be false. While something of the attitude implied by the cliché may remain, Popper’s original point about the logical structure of scientific discovery has difficulty standing up to scrutiny.
In a series of famous works starting in the late 1950s, Popper criticized some (supposedly) scientific fields of study as insufficiently rigorous. It seemed to him that some researchers were focused only on finding positive evidence that could be used to confirm their favorite theories rather than really challenging their theories by trying to find evidence against those theories. For example, Freudian psychologists frequently claimed scientific success after showing that Freudian theory was able to explain a wide range of proposed human behaviors. However, as Popper pointed out, we should be suspicious of that supposed success after recognizing that the theory is so vague and malleable that it can be bent to explain any conceivable human behavior.
Popper labeled such theories “unfalsifiable” and argued that a properly scientific theory should instead tell us what ought not happen. If repeated attempts to find the theory’s forbidden phenomena all fail, then, and only then, has the theory passed a truly risky test and earned some scientific praise. Freudian theory was not capable of subjecting itself to that sort of risky testing, and so was impossible to reject, thus rendering it unscientific according to Popper.
Popper’s views on science were guided by his preference for formal logic. Using particular instances of positive evidence to support a general conclusion, i.e., moving from the particular to the general, requires the use of inductive logic. Unfortunately, it has long been understood that induction can never conclusively prove a general statement about nature to be true. On the other hand, however, when we use negative evidence to contradict a general statement, i.e., when we falsify, we are using deductive logic, and unlike induction, deduction can provide conclusive proof. Thus we arrive at the cliché quoted at the beginning of the essay.
Popper understood that in order for falsificationism to be an accurate account of scientific reasoning, it must describe actual scientific practice. With that in mind, Popper picked the famous Eddington experiment of 1919 in which starlight was observed to follow a curved path around the sun. Newton’s long-standing theory of physics made the general claim that light never follows a curved path through a vacuum, yet that exact curving phenomenon was observed. According to Popper, that observation alone was enough to falsify Newtonian theory, allowing Einstein’s general relativity to take its place.
If Popper’s description of scientific reasoning were correct, then the 1919 episode would indeed be powerful support for falsificationism. However, it turns out that Popper’s description didn’t fully capture scientific practice. Rather than reject Newtonian theory outright, the scientific community defended their familiar and successful older theory. It was suggested that due to their limited measuring abilities at the time, it was entirely possible that the sun’s corona extended out far enough to refract the light as it passed. It wasn’t until more careful and reliable work continued to support Einstein over Newton that the scientific community very gradually shifted to general relativity. In Popper’s defense, one could claim that the Newtonians were just being stubborn, and if they had followed proper scientific logic, they would have rejected their old theory and stopped trying to come up with wild ways to defend it. To see why this is a mistaken characterization, let’s look at some more examples of unquestionably good and unquestionably bad scientific practice.
Suppose a student in Chem 101 is conducting a laboratory exercise in which the liquid in a test tube is supposed to turn blue. Instead, the liquid turns green and the student, following Popper’s reasoning, claims to have falsified the current theory of chemistry. That obviously is bad science because the conclusion is far too hasty. The more reasonable explanation is that the experimenter did something wrong. And that’s not just true for beginners. After bringing the Large Hadron Collider up to full power for the first time, the scientists at CERN failed to find the Higgs boson. If they followed Popper, they would have concluded that the standard model of quantum mechanics was false and stopped looking. Again, that would have been be far too hasty. Instead, they decided to keep searching until they could no longer blame the search’s failure on problems with their methods or equipment, and their efforts eventually paid off in grand fashion.
Popper’s main problem is that his deductive process of falsificationism can never provide a clear refutation of a theory. There always is the possibility that the theory is correct and it was some other detail of the experiment that was responsible for the negative outcome. He may have been right to insist that scientific theories should be subjected to risky tests, but Popper went too far in insisting that the practice of science is a clear-cut deductive process of elimination.
Perhaps, then, the cliché cited at the beginning of this essay should be amended to say, “A million successful experiments cannot prove a theory correct, but one failed experiment can prove that either the theory is wrong or that some mistake was made in the experimental procedure or something totally unexpected happened.” That’s a much messier and more confusing way to describe the scientific process, but nature itself is messy and confusing, so perhaps we should not expect our investigation of it to be much different.
About the Author
Mike is a philosophy instructor at the University of Colorado, Boulder. He earned a bachelor’s degree in biology in 1993 and a master’s in biology in 1995. He then taught high school science classes for several years before going back to grad school at the University of Colorado to earn his Ph.D. in philosophy in 2011. Unsurprisingly, he specializes in philosophy of science and philosophy of biology. Mike also enjoys the classic summertime Colorado activities like hiking, biking, camping, gardening, and going to bluegrass music festivals.