Occam’s Razor and the Virtue of Science

Occam’s razor propels us towards explanations, but it cannot confirm them. That is why science exists. It indiscriminately tells us whether ideas work or not.
“Occam’s razor, or the principle of parsimony, tells us that the simplest, most elegant explanation is usually the one closest to the truth.”

What happens when four years of your lab’s work, a beautiful hypothesis, and an explanation lauded by your scientific peers is “crushed by data”? What do you do? Even greater, what is your role in accepting and communicating a reality ostensibly pinned against you? This is the story of Dr. Steve Nowicki, Professor of Biology at Duke University, whose lab investigates animal behavior in relation to evolutionary fitness.

In the early 2000s, the Nowicki lab was piecing together a puzzle—the puzzle of songbird mating behavior. At the time, it was known that female songbirds used attributes of male birdsong to choose their mate. However, no one fully understood how or why this happened from an evolutionary angle.

In trying to better uncover this arena, Dr. Nowicki’s lab made several discoveries related to the function of birdsong: 1) that not all males learned birdsong equally, 2) females seemed to display a preference for better-learned songs than worse ones, and 3) males that faced early stress during development had lower cognitive ability and did not learn songs as well.

As these different constituent understandings of mating behavior and birdsong vocalization began to coalesce, a beautifully simple hypothesis seemed to take shape. “Developmental stress, song learning, and cognition,” published in 2014 by the Nowicki lab, proposed that female songbirds chose males with better-copied birdsongs because males with the better songs were smarter.

As Nowicki would describe it, the explanation was perfectly “elegant” and just “made sense.” It was realistic, evolutionarily, and biologically sound, and fit the mold of evidence that currently existed. The hypothesis quickly gained traction, with Nowicki’s peers and science media outlets describing it as a gripping revelation. Nowicki’s lab flourished as it never had before.

But soon, “dark clouds” started to rise on the horizon. Studies showed that certain relationships failed to hold under field conditions, and the link between birdsong and the ability to perform other cognitive tasks was questionable. At this chronological junction between fame and tragedy, Nowicki had a decision to make. He could shun the counterevidence under the drapes of glory, or he could give it credence and test it—potentially forfeiting the theoretical perfection of his idea. With immense courage and ethical grounding, Nowicki chose the latter.

In 2017, Nowicki published his findings in the journal Animal Behavior: “Song is not a reliable signal of general cognitive ability in a songbird.” Each written word pierced deeper and deeper into the once magnificent hypothesis: among songbirds that were bred, taught song, then cognitively assessed in a controlled laboratory environment, none of “ 15 correlations linking song measures with cognitive measures were [statistically] significant.” There was no evidence showing that males with better birdsong were any smarter.

Nowicki’s latest research had disproved the culmination of four years of hard work, an idea he had nurtured and brought to life. It was crushing. But Nowicki recognized he had to tell the world.

Science pushes us towards perfect conclusions—the conclusions we want to hear. It rewards positive associations, discoveries, and novel ideas. Even more, it pressures us to achieve those things. Journals, universities, and funders don’t want to hear about a therapy or trial that didn’t work; they want to invest resources in ones that did.

In an enterprise that has become so competitive and glory-seeking, it becomes easy to undermine the bedrock of what science hopes to achieve at all. It becomes tempting to capitulate the human grounding, empathy, and ethical commitment needed to help us accurately understand our world, in favor of personal triumph. Dr. Nowicki made the decision to rethink his own creation, and then he made the decision to dismantle it in front of the world. He acted to preserve the rectitude and integrity of science. He acted with nobility and character that every one of us should strive to attain.

Dr. Nowicki concluded his seminar with a quote by Claude Bernard: “We must never make experiments to confirm our ideas, but simply to control them,” or as Dr. Nowicki likes to say, “simply to test them.” Nowicki’s story reminds us that science, at the end of the day, is a very human enterprise, riddled with hope, ambition, grief, tragedy, success, and failure. This is the side of science we are often shielded from.

However, parallelly, we cannot let science become a slave to human sentiment. Science needs objectivity to mean anything. While principles of philosophy—like Occam’s razor—can help us identify a likely explanation, only science can tell us whether an idea truly holds or not.

On the basis of this understanding, we all have an important responsibility toward the scientific enterprise, a responsibility unpinned throughout the entirety of Dr. Nowicki’s talk: to safeguard the foundational purpose, utility, and virtue of science across our own work, and never compromise it for ourselves.


Ronit Sethi, Huang Fellow ’24

Sethi_RonitRonit is an undergraduate student from Princeton, NJ pursuing majors in biology and global health.