Equations for traffic and marriage

One safe prediction we can make about the coming election season is that there will be an abundance of pundits with conflicting opinions. We'll hear from TV personalities and party functionaries. But, if science writer Philip Ball is right, the people we really should listen to are physicists.

In "Critical Mass: How One Thing Leads to Another," Ball gives a sprawling account of physics over the past several centuries - from thermodynamics to complexity theory - showing how fundamental insights about the behavior of matter can be adapted to understand the dynamics of society. He applies this claim intriguingly to a variety of social, economical, and political situations, showing, for instance, that voting follows patterns akin to magnetization, and marriage rates resemble the behavior of gasses and liquids.

While neither cohesive nor gracefully written, the book provides a solid foundation for assessing a compelling argument: The tools of statistical physics cannot reasonably be ignored by a government seeking to construct viable social policy.

"Once we acknowledge the universality displayed in the physical world," Ball writes, "it should come as no surprise that the world of human social affairs is not necessarily a tabula rasa, open to all options." In other words, before deciding how we ought to act, we should figure out what can - and can't - be done.

Of course, Ball isn't the first to suggest that societies might learn from science. In the 17th century, Thomas Hobbes famously built a political system around the rational study of what he believed to be human nature. "The skill of making, and maintaining Commonwealths consisteth in certain Rules, as doth Arithmatique and Geometry," Hobbes claimed in "Leviathan" (1660). With those rules, he deftly extrapolated from man's insatiable desire for power, the need for absolute monarchy.

Ball rightly recognizes Leviathan's importance: "[Hobbes] does not describe a society ready-made and shaped by his own preferences, but builds it up, with careful logic, from his mechanistic view of how humans behave."

Yet he also rightfully dismisses Hobbes's draconian conclusions. How? Not only by rejecting his questionable premise, but also, more significantly, by challenging his methodology.

Just as one can't explain the behavior of planets by studying the motion of atoms, Ball believes that one cannot understand how societies work by examining the motivations of individuals. "One of the features of collective behavior arising from local interactions," he writes, "is that it becomes impossible to deduce the global state of a system purely by inspecting the characteristics of individual components." Even if people individually are power hungry, we can't safely assume that civilization as a whole works that way.

On the contrary, we must recognize that "a society creates its leaders, its customs, its fashions, and its problems through a mass of mutually interdependent decisions."

Statistical physics was developed to analyze such situations, and, while initially used to model the behavior of gasses, its tools have proven equally applicable to crowds. That is, people in the aggregate seem to follow the same basic physical laws as matter.

For instance, Ball notes that statistical graphs of the way neighborhoods suddenly turn to crime show an uncanny resemblance to the way gases suddenly turn to liquids. (See box below.)

In a close-knit neighborhood, worsening social conditions may lead to very little increase in criminal behavior, just as a gas may not transform into a liquid even under increased pressure. But at some point, an additional worsening of social conditions suddenly leads to a dramatic increase in crime, just as, under the right conditions, an additional increase in pressure can suddenly convert a gas into a liquid.

From this it follows that crime rates are prone to extreme change only at particular points. Understanding this process could allow policymakers to allocate resources more intelligently for maximum effect.

Similarly, modeling traffic jams on a computer, researchers have found the automotive equivalent of "phase transitions" from liquid to solid. Just as liquids can be supercooled below the freezing point without seizing up if the process isn't rushed, large numbers of cars can travel a road uncongested if nobody abruptly accelerates or breaks. Such insights could lead to better designs for entrance and exit ramps.

Nevertheless, Ball is determined not to transform statistical physics into a new form of social Darwinism. He argues that science should be "a servant and guide, not a dictator." In other words, intellectually powerful as science may be, it carries no moral authority.

While he's absolutely right, he's perhaps naive to expect others to heed his warnings in a society that routinely engineers ethics to suit its technology. More deeply troubling, however, is his unmitigated embrace of pragmatism: "By gaining [sufficient scientific] knowledge," he writes, "we might hope to adapt our social structures to the way things are rather than the way some architect or politician or town planner thinks they ought to be." For architecture, that would spell the end of decoration and whimsy. For society, it would mean giving up our dreams.

• Jonathon Keats is a board member of the National Critics Circle. He reviews books from San Francisco.