Today, we ask if there was life before Clovis. The University of Houston's College of Engineering presents this series about the machines that make our civilization run, and the people whose ingenuity created them.

Scientists are always at their worst when they're most certain. Science looks the way it's supposed to look only when it faces mystery. When Dirac and de Broglie struggled with particles and waves in the 1920s, that was real combat with mystery.

It was just then, in the '20s, that another piece of common wisdom began falling apart. In 1920, we believed the age of human life in the Americas was only a few thousand years. Then, in 1926, we found remains at a site near Clovis, New Mexico. The site held lots of tools -- and a few bones -- all from the late glacial period. Anthropology riveted our interest as the age of the first Americans suddenly stretched to 12,000 years.

For half a century we turned up other sites of comparable age, but none older. As the years passed, scientific inquiry turned to conviction. Twelve thousand years became a Maginot line that anthropologists could not cross.

Now we're again finding what might be far older sites. And scholars square off to do battle. The result is, to my mind, the game of science at its best. Let's look at the evidence:

Carbon dating tells us that mammoth bones in the Yukon are 15 to 20 thousand years old. The bones seem to have been worked by stone tools, but were they, really? Remains in a Pennsylvania site are 13 to 20,000 years old. They include two human bones whose age seems to be 12 to 14 thousand years. But skeptics find ways to question the carbon dating.

We've found arrowheads among 13,000-year-old mastodon bones in Venezuela. But they might've been dumped there by rising water. Similar doubts cloud similar findings in Chile.

The most notorious find of all is a recent one in Brazil. An arrangement of charcoal could be a campsite. It gives carbon dates over 30,000 years old. But was it really a campsite?

The pervasive problem is the lack of a corpus delicti. Where's the body? Until we find human remains, it's easy to be skeptical. The trouble with experimental proofs is, they aren't absolute. Defenders of the old Clovis dates fall back on a legal definition of proof, not a scientific one. Instead of accepting the most likely conclusion, they demand that we erase all reasonable doubt. And that's something science can never do.

So we draw battle lines. So much scholarship has rested on the Clovis dates. The new sites threaten whole lifetimes of intellectual commitment. But, however the contest should end, New-World anthropology is back in the public domain. The game is afoot again; and for that, knowledge is alive.

I'm John Lienhard, at the University of Houston, where we're interested in the way inventive minds work.

(Theme music)

Marshall, E., Clovis Counterrevolution. Science, Vol. 249, 17 August 1990, pp. 738-741.

Morell, V., Confusion in Earliest America. Science, Vol. 248, 27 April 1990, pp. 439-441.