Certainties of Very Low Probability

Oxymorons are paradoxes gone bad. We learn simple causal relations with great facility. As Clark Glymour has observed, even infants know this: to get the pacifier, tug on the blanket it lies on. But we have a harder time getting our minds around the meaning of very small probabilities over very long time scales, likely because our hominid ancestors never lived long enough to have to worry about them.

Some years ago, when I was in graduate school at Northwestern University, we had a Friday noon seminar on aspects of science and technology. I recall one meeting that featured a paper on high-energy physics. One of the senior physicists, Arnold Siegert, who had been a student of Heisenberg’s, got into an argument with a charming but not overly reverent mathematician (whose name I regrettably don’t remember) about the physical interpretation of mathematical formalisms. It was the ancient quarrel about whose reality was better, the mathematicians’ Platonic idealism or the physicists’ experimental materialism. At issue was some particle-decay process with a probability of ten to the minus something or other, which, when translated, meant that it should occur approximately once every thirty or so billion years, or once in twice the believed age of the universe. Finally, the mathematician, in some exasperation, asked, “Arnold, what would you do if you observed this phenomenon?” There was a silence. Then Siegert replied, almost impishly, “Not tell anyone.” It’s a splendid stratagem for improbable events in subatomic physics, but it doesn’t scale well for hurricanes. Too big.

What’s striking about Hurricane Katrina is that, like the nuclear physicists, everybody knew it was going to happen—sooner or later. Certainly anyone who’d lived in New Orleans or in Louisiana or on the Gulf Coast for very long knew. My grandmother was born and raised in New Orleans, in what’s now more hopefully than veridically called the Garden District, and I grew up with the folklore of the River and hurricanes. Later, when I was at Tulane, we lived in an apartment building that faced South Claiborne Avenue but backed up to the old Sugar Bowl. Between our building and the stadium were the university’s practice fields, five or six of them. On each side of each regulation 100-yard field were two or three gently contoured, sodded sumps, perhaps 30 yards long and 10 yards wide, maybe 4 or 5 feet deep, with a large drain at the bottom. When the cloudbursts came, which was often, the sumps would rapidly fill up. South Claiborne is about 60 yards wide at that spot, with two lanes in each direction separated by a broad grass esplanade, which covers one of the several major canals that drain New Orleans. From our sixth-floor apartment we could see the end of South Claiborne as it curved around, and we would watch for the pumps down there to come on. Sooner or later we’d see the smoke from the big diesel engines that drove the pumps, and pretty soon the water in the sumps would go down, usually in only ten or fifteen minutes. When it was later rather than sooner, there wasn’t much mystery about what would happen if it were ever never.

I also remember one blurry morning sitting in Jackson Square, as gentle night gave way to merciless day, watching the strippers go home, tired and flat-footed. I had one of those “what’s wrong with this picture?” moments as it dawned on me that the hull of the freighter I was looking up at was visible above the top of the levee. We clambered across the railroad tracks behind the old Jax brewery (in those less troubled days the night watchman only fussed at us a little) and up onto the levee. Sure enough, the river stage was a good 20 or 25 feet higher than where I’d been sitting in Jackson Square.

The river levees at New Orleans are challenged, severely, at least once a year, usually more often, and the powers that be have quite rationally devoted the most attention to them.{1} The rest of the city is protected too, but driving around in the then recently built subdivisions out by City Park and Lakefront Airport, on Lake Pontchartrain, it was pretty obvious what the consequences would be if the northeast quadrant of a “perfect storm” got into Chandeleur and Mississippi Sounds: a massive storm surge would be driven through Lake Borgne into Lake Pontchartrain, and then into New Orleans. Sooner or later.

Some years later, I drove over to Old River, where the Red River, the Mississippi, and the Atchafalaya come together, about 300 river miles above New Orleans. Old River Control, so memorably portrayed in John McPhee’s Control of Nature, is a set of massive spillways, a gargantuan navigation lock, and now a hydroelectric power station (then under construction) that together regulate the flow of water from the Mississippi into the Atchafalaya Basin. The entire complex was nearly obliterated in the great 1973 flood. The basic problem is that the River—the Mississippi, anthropomorphized since the beginning of time—“wants” to go down the Atchafalaya; the distance to the Gulf is about one-third, and the gradient is twice as steep. We don’t want it to.

I stood on the lock structure, at least half a mile long, talking to the solitary lock-tender, a native Cajun. The lock was so big—to admit the huge tows that pass between the rivers—that he used an old Schwinn bicycle to commute between the gate-control houses at either end of it. We were talking about the River and what had happened in 1973. I noticed a big island off to the northwest and asked if it were Turnbull Island, allowing that my granddaddy had logged it in 1910. That and the fact that I’d said “Turnbull” correctly (think about how Creedence Clearwater Revival says “turnin’” or “burnin’” in “Proud Mary”—it’s a pronunciation unique to about four parishes in central Louisiana) put him at ease. I wasn’t some outsider come to make trouble, dinky little foreign car or not, but a near-native, come home to talk about the River. We chatted for a while, and I finally asked him straight out whether he thought the works at Old River would hold. He laughed and said something to the effect that the River had been doing what it wanted for a lot longer than we had been trying to control it, and “sooner or later . . . .” I asked what he was going to do when it happened. He laughed again and, pointing to where he stood on the massive concrete lock structure, said, “I’m gonna come right here. It’s the highest point in eight parishes.”

What’s curious about these little vignettes is that they’re all implicitly cast as narratives about nature, when they’re really as much about what Tom Hughes calls “the human built world.” We do live on a restless, if not malevolent, planet, whose dynamics we barely grasp. But we also do stupid things. Short-run avarice always trumps long-run prudence, and so we build taxpayer-insured houses on barrier islands and in known floodplains. Our interventions have unintended, if not unforeseeable, consequences. The land inside the levees, along the rivers, is higher now than the land outside, which the levees ostensibly protect, as the land outside, deprived of sediment, slowly subsides.

These vignettes also have something else in common: an attitude toward certainties of very low probability. It’s not bravado or fatalism; too melodramatic. And it’s not a “form of life” or “a life world”; too highfalutin. It’s just there. People in New Orleans and south Louisiana know about the River and about hurricanes the way people in other places know about earthquakes or volcanoes or mudslides or the lethal build-up of carbon dioxide in deep lakes. It’s part of natural-born culture, like the way “ur” is pronounced in four parishes in Louisiana. It doesn’t diminish the human (and animal) tragedy of a Katrina to say that everybody knew. What historians of technology might want to think about is how we think about what’s ours and what’s nature’s, why we conflate the two, and whether our tacit intuition that they are the same speaks more wisely than we know.

{1}Rational in this sense: The expected value of a probabilistic bad event is its probability times its cost. A risk portfolio comprises the sum of the risks’ expected values. If resources can be applied to reduce probabilities or ameliorate costs, the value of the entire portfolio (total risk) is minimized when the marginal decrement in expected value per dollar spent is equal for all risks. Thus, assuming it’s immaterial whether New Orleans is flooded by the river or by a storm surge, it is rational to expend disproportionate resources to combat the more likely risk, which is river flooding.

Ironically, publicly funded flood-control measures more often than not have contradictory results. Usually they are undertaken explicitly as economic development initiatives and, if successful, as in New Orleans’ eastern wards, increase the cost of a disaster at the same time they putatively reduce its probability. Thus expected value is little improved. Moreover, implementation of this rationalist risk-management strategy assumes perfect knowledge and perfect foresight (known probabilities and costs), which is usually transmogrified into the assumption that the future will be pretty much like the past. Thus economics, that allegedly most rigorous of social-science disciplines, routinely makes an assumption that few if any professional historians would defend unequivocally.