I've been mulling over the potential, and limits, of social science again lately. One reason is that last month philosopher James Weatherall of the University of California at Irvine visited my school, Stevens Institute of Technology, to talk about his new book The Physics of Wall Street. Weatherall, who has a Ph.D. in physics as well as in philosophy, argued that the methods of physics can help make economics more rigorous. Then someone sent me an article in The Economist on how "data from social networks are making social science more scientific."

Goaded by these optimistic claims, I decided to post thoughts about social science that I presented two years ago in The Chronicle of Higher Education. From its inception, social science--which includes economics, sociology, anthropology, political science and social psychology—has struggled for respect. In the early 19th century, French philosopher Auguste Comte proposed a scientific hierarchy ranging from the physical sciences at the bottom up through biology to the "queen" of sciences, sociology, at the top. A science of human social behavior, Comte contended, could help humanity make moral and political decisions and construct more efficient, just governments.

Comte, who spent time in a sanitarium for mental illness, had admirers--notably John Stuart Mill--but he was viewed by many of his contemporaries as a crank. He died in 1857 without ever landing a full-time university post or indeed any steady employment. Today, social science receives much less federal funding than the biological and physical sciences do. Social scientists are accused of being "soft," of trafficking in theories so lacking in precision and predictive power that they don't deserve to be called scientific.

Some social scientists—I'll call them "softies"—shrug off this criticism, because they identify less with physicists and chemists than with scholars in the humanities. Stevens Institute is a case in point: Social science falls within the jurisdiction of the Stevens College of Arts & Letters, which also encompasses philosophy, history, literature, music and my own humble discipline, science communication. As far as I can tell, my social-science colleagues aren't seething with resentment at being lumped together with the humanities folks.

Other social scientists, "hardies," yearn for and believe they can eventually attain the same status as, say, molecular biology. Softies and hardies have been fighting for as long as I can remember. In 1975, for example, the Harvard biologist E.O. Wilson contended in his blockbuster Sociobiology that social science would only become truly scientific by embracing evolutionary theory and genetics. Horrified softies denounced sociobiology as a throwback to social Darwinism and eugenics, two of the most noxious social applications of science.

The term "sociobiology" became so controversial that it is rarely used today, except by softies as an insult. Hardies nonetheless embraced the tenets of sociobiology. They tacked the term "evolutionary" to their fields—spawning disciplines such as evolutionary psychology and evolutionary economics—and churned out conjectures about the adaptive origins of war and capitalism.

More recently, as the prestige of neuroscience has surged, hardies have discovered the benefits of including magnetic-resonance imaging and other brain-scanning experiments in grant proposals, and they have attached the prefix "neuro" to their disciplines, yielding coinages such as neuroeconomics and neuroanthropology.

Hardies also emulate the hardest science of all: physics. Thus we now have econophysics, which models economic activity with concepts borrowed from fluid dynamics, solid-state physics and statistical mechanics. (For a terrific overview, see the aforementioned The Physics of Wall Street.) This alliance has especially deep roots: Comte sometimes used the term "social physics" in lieu of sociology. But modern researchers, unlike Comte, can run their complex mathematical models on powerful computers.

Softies look askance at the aspirations of hardies—with good reason. The recent recession provides a powerful demonstration of social science's limits. The world's smartest economists, equipped with the most sophisticated mathematical models and powerful computers that money can buy, did not foresee—or at any rate could not prevent—the financial calamities that struck the United States and the rest of the world in 2008. As philosopher Paul Feyerabend once said: "Prayer may not be very efficient when compared to celestial mechanics, but it surely holds its own vis-à-vis some parts of economics."

Even when fortified by the latest findings from neuroscience, genetics, and other fields, social science will never approach the precision and predictive power of the hard sciences. Physics addresses phenomena—electrons, elements, electromagnetism, the nuclear forces, gravity—that are relatively simple, stable and amenable to precise mathematical definition. Gravity works in exactly the same way whether you measure it in 17th-century England or 21st-century America, in Zambia or on Alpha Centauri. Every neutron is identical to every other neutron.

In contrast, the basic units of social systems—people—are all different from each other; each person who has ever lived is unique in ways that are not trivial but essential to our humanity. Each individual mind also keeps changing in response to new experiences—reading Thus Spoke Zarathustra, watching Lord of the Rings, banging your head on the ice while playing pond hockey, having a baby, teaching freshman composition. Imagine how hard physics would be if every electron were the unique product of its entire history.

Societies also vary markedly across space and time. France in 2013 is radically different than it was in Comte's era. The United States today is quite different than it was a century, a decade or a year ago. Social scientists are chasing a moving target, one they can never catch. As anthropologist and archetypal softy Clifford Geertz once wrote, social scientists can construct only "hindsight accounts of the connectedness of things that seem to have happened: pieced-together patternings, after the fact."

Here is the biggest difference between social and hard science: Protons, plasmas and planets are oblivious to what scientists say about them. Social systems, on the other hand, consist of objects that watch television; listen to the radio; read newspapers, journals, books, and blogs; and consequently change their behavior. In other words, social-science theories can transform societies if people believe in them.

Even Comte made his mark. His writings inspired the founders of the republic of Brazil and the motto on the nation's flag: "Ordem e Progress" (Order and Progress). More significantly, Comte influenced Marx, whose social theory profoundly altered the course of human history.

So we are left with a paradox: Although social science is in many respects quite weak, it can also be extraordinarily potent in terms of its impact, for ill or good, on our lives. Think of all the harm done in the name of Marx—and of social Darwinist and free-market theorists, from Herbert Spencer to Milton Friedman.

But social scientists can improve the world, too. Those I admire most combine rigorous empiricism with a resistance to absolute answers. These are researchers like anthropologist Sarah Blaffer Hrdy, who examines the behavior of primates and early humans for insights into modern gender roles; economist Jeffrey Sachs, who seeks ways to reduce third-world poverty; or political scientist Gene Sharp, an authority on nonviolent social activism.

Social scientists are especially dangerous when they insist—and convince others—that they have discovered absolute truths about humanity, truths that tell us what we are and even what we should be. Hence social scientists—more than any other scientists—should be humble, or at least modest, in making claims.

Here's a more specific suggestion: Social scientists should consider identifying not with the harder sciences or the humanities but with engineering.

I started my career writing for an engineering magazine, and now I teach at an engineering school, so I know and respect engineers. They don't seek "the truth," a unique and universal explanation of a phenomenon or solution to a problem. In fact, engineers would scoff at such a formulation of their work. They seek merely answers to specific, localized, temporary problems, whether building a bridge with less steel or a more efficient solar panel or a smartphone with a bigger memory. Whatever works, works.

In the same way, social scientists should eschew the quest for truths about human behavior. They should instead focus more intensely on finding answers to specific problems, whether our current economic woes, the inefficiency of our health-care system or our reliance on military force to resolve disputes.

In spite of its weaknesses, social science—when applied wisely—can do even more than the hard sciences to make the world a better place. Comte was right about that.

Photo: Dieter Drescher, Flickr.