But now imagine a totalitarian state, not overrun by enemy troops, but one thriving and self-confident. Imagine a tradition in which orders are obeyed without question. Imagine that those involved in the operation are supplied a cover story: The asteroid is about to impact the Earth, and it is their job to deflect it—but in order not to worry people needlessly, the operation must be performed in secret. In a military setting with a command hierarchy firmly in place, compartmentalization of knowledge, general secrecy, and a cover story, can we be confident that even apocalyptic orders would be disobeyed? Are we really sure that in the next decades and centuries and millennia, nothing like this might happen? How sure are we?

It’s no use saying that all technologies can be used for good or for ill. That is certainly true, but when the “ill” achieves a sufficiently apocalyptic scale, we may have to set limits on which technologies may be developed. (In a way we do this all the time, because we can’t afford to develop all technologies. Some are favored and some are not.) Or constraints may have to be levied by the community of nations on madmen and autarchs and fanaticism.

Tracking asteroids and comets is prudent, it’s good science, and it doesn’t cost much. But, knowing our weaknesses, why would we even consider now developing the technology to deflect small worlds? For safety, shall we imagine this technology in the hands of many nations, each providing checks and balances against misuse by another? This is nothing like the old nuclear balance of terror. It hardly inhibits some madman intent on global catastrophe to know that if he does not hurry, a rival may beat him to it. How confident can we be that the community of nations will be able to detect a cleverly designed, clandestine asteroid deflection in time to do something about it? If such a technology were developed, can any international safeguards be envisioned that have a reliability commensurate with the risk?

Even if we restrict ourselves merely to surveillance, there’s a risk. Imagine that in a generation we characterize the orbits of 30,000 objects of 100-meter diameter or more, and that this information is publicized, as of course it should be. Maps will be published showing near-Earth space black with the orbits of asteroids and comets, 30,000 swords of Damocles hanging over our heads-ten times more than the number of stars visible to the naked eye under conditions of optimum atmospheric clarity. Public anxiety might be much greater in such a time of knowledge than in our current age of ignorance. There might be irresistible public pressure to develop means to mitigate even nonexistent threats, which would then feed the danger that deflection technology would be misused. For this reason, asteroid discovery and surveillance may not be a mere neutral tool of future policy, but rather a kind of booby trap. To me, the only foreseeable solution is a combination of accurate orbit estimation, realistic threat assessment, and effective public education—so that in democracies at least, the citizens can make their own, informed decisions. This is a job for NASA.

Near-Earth asteroids, and means of altering their orbits, are being looked at seriously. There is some sign that officials in the Department of Defense and the weapons laboratories are beginning to understand that there may be real dangers in planning to push asteroids around. Civilian and military scientists have met to discuss the subject. On first hearing about the asteroid hazard, many people think of it as a kind of Chicken Little fable; Goosey-Lucy, newly arrived and in great excitement, is communicating the urgent news that the sky is falling. The tendency to dismiss the prospect of any catastrophe that we have not personally witnessed is in the long run very foolish. But in this case it may be an ally of prudence.

Meanwhile we must still face the deflection dilemma. If we develop and deploy this technology, it may do us in. If we don’t, some asteroid or comet may do us in. The resolution of the dilemma hinges, I think, on the fact that the likely timescales of the two dangers are very different—short for the former, long for the latter.

I like to think that our future involvement with near-Earth asteroids will go something like this: From ground-based observatories, we discover all the big ones, plot and monitor their orbits, determine rotation rates and compositions. Scientists are diligent in explaining the dangers—neither exaggerating nor muting the prospects. We send robotic spacecraft to fly by a few selected bodies, orbit them, land on them, and return surface samples to laboratories on Earth. Eventually we send humans. (Because of the low gravities, they will be able to make standing broad jumps of ten kilometers or more into the sky, and lob a baseball into orbit around the asteroid.) Fully aware of the dangers, we make no attempts to alter trajectories until the potential for misuse of world-altering technologies is much less. That might take a while.

If we’re too quick in developing the technology to move worlds around, we may destroy ourselves; if we’re too slow, we will surely destroy ourselves. The reliability of world political organizations and the confidence they inspire will have to make significant strides before they can be trusted to deal with a problem of this seriousness. At the same time, there seems to be no acceptable national solution. Who would feel comfortable with the means of world destruction in the hands of some dedicated (or even potential) enemy nation, whether or not our nation had comparable powers? The existence of interplanetary collision hazards, when widely understood, works to bring our species together. When facing a common danger, we humans have sometimes reached heights widely thought impossible; we have set aside our differences—at least until the danger passed.

But this danger never passes. The asteroids, gravitationally churning, are slowly altering their orbits; without warning, new comets come careening toward us from the transplutonian darkness. There will always be a need to deal with them in a way that does not endanger us. By posing two different classes of peril—one natural, the other human-made—the small near-Earth worlds provide a new and potent motivation to create effective transnational institutions and to unify the human species. It’s hard to see any satisfactory alternative.

In our usual jittery, two-steps-forward-one-step-back mode, We are moving toward unification anyway. There are powerful influences deriving from transportation and communications technologies, the interdependent world economy, and the global environmental crisis. The impact hazard merely hastens the pace.

Eventually, cautiously, scrupulously careful to attempt nothing with asteroids that could inadvertently cause a catastrophe on Earth, I imagine we will begin to learn how to change the orbits of little nonmetallic worlds, smaller than 100 meters across. We begin with smaller explosions and slowly work our way up. We gain experience in changing the orbits of various asteroids and comets of different compositions and strengths. We try to determine which ones can be pushed around and which cannot. By the twenty-second century, perhaps, we move small worlds around the Solar System, using (see next chapter) not nuclear explosions but nuclear fusion engines or their equivalents. We insert small asteroids made of precious and industrial metals into Earth orbit. Gradually develop a defensive technology to deflect a large asteroid or comet that might in the foreseeable future hit the Earth, while, with meticulous care, we build layers of safeguards against misuse.

Since the danger of misusing deflection technology seems so much greater than the danger of an imminent impact, we can afford to wait, take precautions, rebuild political institutions—for decades certainly, probably centuries. If we play our cards right and are not unlucky, we can pace what we do up there by what progress we’re making down here. The two are in any case deeply connected.