Dangers without an author

"You can't see radon. And you can't smell it or taste it." That is the opening of the US Environmental Protection Agency's guide to the gas, which seeps out of the ground into basements. "Testing is the only way to know."¹ The agency estimates that radon causes about 21,000 lung cancer deaths a year in the United States, and calls it the second leading cause of lung cancer after smoking.²

Few people, as far as anyone can tell, lie awake over radon; nothing read for this piece measures how much it is feared. The idea behind this piece is that this is no accident: fear follows the eyes. The dangers that frighten people have a look, a shark, a crash, a fire. Many of the dangers that harm people have none.

That idea is half right, and the half that is wrong is the interesting part.

Fear needs a cue

Start with how fear could work at all. It is a detector, and a detector needs something to detect: a movement, a sound, a wound, an attacker. For most of human history those cues were reliable signs of danger. A hazard that gives off no cue at all, no color, no smell, no moment, should leave the detector silent.

Radon fits. It cannot be seen, smelled or tasted.¹ The harm it does, if any, shows up as lung cancer, which looks like any other lung cancer.

The problem: radiation

But radiation fits too, and radiation is anything but ignored.

When the psychologist Paul Slovic mapped how people judge hazards, in a 1987 Science paper (known here through a summary of it), two clusters of traits explained most of it. One was dread: catastrophe, lack of control, fatal outcomes. The other was the unknown: risks that cannot be observed, whose effects are delayed, that the people exposed may not know about.³ Nuclear power sat at the extreme of both, feared far beyond what experts estimated.³

So invisibility does not silence fear. Radiation and radon are equally impossible to see, smell or feel. Radon is a radioactive gas, from the natural decay of uranium in the ground.¹ It is a product of the same physics as nuclear power. One is among the most dreaded hazards of the last century; the other, on this piece's premise, barely registers.

The difference is not what the eye can see. It is whether the danger has an author.

Authored and authorless

Radiation came with a story: a bomb, an industry, accidents with place names, someone who decided and someone to blame. Radon comes out of the ground. Nobody put it there. There is no event, no villain and no day when it happened.

Fear does not need a danger to be visible. It needs a cue, and an invisible danger can borrow one: from a story, which supplies an agent and a moment, or from an instrument, which supplies a reading. The EPA puts it plainly: "Testing is the only way to know."¹ Without the instrument, there is nothing for fear to attach to.

The dangers fear misses, on this account, are the ones with no author and no instrument: natural, slow, and quiet.

When the visible danger pushes you into the invisible one

The clearest case of fear choosing the wrong danger involves a risk that is not invisible at all.

After September 11, 2001, Americans flew less and drove more. The psychologist Gerd Gigerenzer compared fatal road crashes in the last three months of 2001 with the same months in previous years and found an excess that, he estimated, corresponded to about 353 deaths, more than the 266 passengers and crew killed on the four hijacked flights.⁴ His figure is an estimate from national totals, not a tally of individual choices, and he presents it as one.⁴

A car is easy to see. What was invisible was the arithmetic. The hijacked planes supplied a cue as strong as any in modern memory; the extra road deaths were spread across a country and a season, each one an ordinary crash on an ordinary day. The danger fear missed had no author and no moment.

Uncounted as well as unfeared

Here the idea turns back on the numbers.

The 21,000 radon deaths are not a count. Nobody's death certificate says "radon." The figure comes from risk models built on studies of miners, working-age men exposed to doses often far higher than homes deliver, extrapolated to the whole population.⁵ The EPA's own uncertainty analysis says its estimates may be accurate "to within factors of about 2 or 3."⁶ A 2017 study argued that diesel exhaust in the mines had been mistaken for radon's effect and put the toll 9 to 26 percent lower, at roughly 15,600 to 19,300 deaths a year.⁵

The agency and its critics agree radon kills thousands of Americans a year. They disagree about how many, and nobody can settle it by counting. That is not a failure of the people involved. A danger with no event leaves nothing to count; it can only be modeled.

So the authorless dangers are under-feared and under-counted, and for the same reason: nothing happens that anyone can point to. The detector in the head and the statistics office share the blind spot.

The strongest objection

The strongest objection is to the whole frame. "People fear the wrong things" is a sentence with a history. In the comments under one summary of Slovic's work, a public-health teacher described seeing it used by companies to dismiss communities' fears of chemical exposure as unscientific over-reaction.³

The objection has force. Fearing an authored danger more than a natural one may not be a mistake. A danger with an author can be stopped by stopping the author, and someone can be held to account. Dread of nuclear accidents is also dread of institutions that might lie about them. Measured by deaths alone, that fear looks miscalibrated. Measured by what fear is for, which includes accountability, it may not be.

That narrows the claim. The point is not that people fear the wrong things. It is that one class of danger, the authorless kind, slips past both fear and counting, and nobody is assigned to notice it.

An open edge

One further implication, which nothing read here tests.

The dangers people could see are the ones they learned to guard against: rails on cliffs, guards on machines, laws against weapons. Seeing a danger made it possible to fight it. If so, the dangers still harming people in large numbers should be, increasingly, the ones that were never seen.

That would mean fear looks worse calibrated over time not because it got worse, but because it already won the fights it was built for. What remains is a world whose biggest dangers have no look, no author and no moment, and are known, when they are known at all, through an instrument and a model.