Why It Didn’t Happen : The Limits of Rationality in Preventing Nuclear War
Orradre Professor of Economics
With his death approaching in 1957, the great mathematician John von Neumann confided to a friend that he was “absolutely certain” there would be a nuclear war that would end all human life. He wasn’t alone in that thinking. Soon afterward, the British novelist C.P. Snow told the New York Times that without massive and speedy disarmament, it was a “mathematical certainty” there would be a nuclear war.
Von Neumann, Snow and many others turned out to be wrong (at least so far), and in a paper just completed, “Schelling, Irrationality and the Event That Didn’t Occur,” Alexander J. Field looks at some of the reasons. Field, the Michel and Mary Orradre Professor of Economics, concludes that mathematical theory failed in predicting the answer to the greatest life-and-death question of the twentieth century.
“When you look back at the Cold War period, it’s surprising how widespread and unrelenting was the pressure to launch a nuclear first strike,” Field says. “Even a pacifist like Bertrand Russell pushed for it; the logic was so overwhelming it would happen.”
It’s the failure of that logic to predict events that interests Field, and he argues in considerable detail that the economist and Nobel Prize winner Thomas Schelling, failed when he tried to buttress the case for deterrence using game theory, because game theory provides a more compelling rationale for a first strike.
Game theory assumes that the players (in this case, the nuclear superpowers) will both behave rationally in the sense of acting efficiently to advance their material self-interest. It might be rational for you not to attack because you feared retaliation, but this requires attributing irrationality to your opponent who must in turn, if deterred, attribute some irrationality to you. One ends up assuming that you and your opponent are both rational and irrational — a theoretically incoherent position, but one that might have some truth.
Followed ruthlessly to its conclusion, game theoretic reasoning offers little support for holding off until attacked and an overwhelming case for first strike. Von Neumann himself believed this, arguing, with respect to the Soviets, “If you say why not bomb them tomorrow, I say why not today? If you say today at 5 o’clock, I say why not one o’clock.”A rational opponent should be attacking you already and in any event would not retaliate if you hit first.
In fact, Field notes, during the Cold War there were ongoing concerns on the part of national security planners about whether a president would have the will to retaliate after a first strike by the other side when the retaliation would serve no purpose other than vengeance.
What instead seems to have overridden the logic for a first strike and allowed mutual deterrence to work were more human factors: We do often hesitate before striking first and we are ready to hit back. “The failure accurately to predict is not due to a defect in the logic,” Field writes in his paper. “The problem is with the implied behavioral assumptions.”
Field argues that within the politics of academia there is a premium placed on using game theory to explore real-world problems because it presumably brings a more analytical focus to the discussion. The problem is that in some cases, such as analyzing nuclear deterrence, the approach doesn’t yield useful results because of what it leaves out.
“The bottom line of this paper,” Field says, “is that nuclear war has been avoided not because we were rational, but because we were human.”