AA Guns and the Birth of Cybernetics

In the 1930s the UK had the following bombing defence policy: “The bombers will always get through.” Which, in my opinion, is a perfectly wonderful stiff-upper-lip attitude to have. Then radar was invented and it became possible to track planes. Actually shooting them turned out to be stunningly difficult.

Creative Commons License photo credit: toolmantim

To hit a plane with an AA gun, you need to aim at where the plane is going to be after your bullets have travelled to reach it. This is hard enough when the plane is merely far away and travelling very quickly. It’s near-impossible when that same plane is actively trying to get out of the way of your bullets.

The answer is cybernetics.

I’ve already written about governors on steam engines which are essentially machines that turn steam into information. The trick is getting the engine to tell the valve how much energy it’s pushing out, which in turn allows the valve to open or close and generally wrestle the boiler into giving a constant flow of energy. The governor is the mechanism by which the boiler talks to the valve. Steam is siphoned off, turned into information and an essentially random output becomes controlled and predictable.

When it comes to shooting at banking and diving planes, a similar insight comes into play. From the perspective of the gun operator, the plane is a random actor; its motions constrained by physics and biology. Given its current location in the sky at one moment, there is a set of points where it is likely to be in the next. Here’s Norbert Wiener solving the problem in 1941:

The resulting flight path was highly irregular but not purely capricious. Those waving lines on the blackboard were old friends of Wiener’s. “There are,” he said, in a rare display of understatement, “in fact, means which will allow one to accomplish the minor task of a quite correct prediction.”

Dark Hero of the Information Age: In Search Of Norbert Wiener–Father of Cybernetics by Flo Conway & Jim Siegelman

Re-conceptualize the situation. It’s not a gun trying to shoot an aircraft. It’s a system trying to ensure that two objects stay in sync. Here’s how the system works: The operator centres the crosshairs on the plane. The predictor works out where the plane is likely to go and points the gun there. The gun fires. The plane moves. The operator adjusts the crosshairs, which corrects the predictor, which moves the gun, which fires the shell.

The plane has no chance. It’s a randomly burbling boiler caught in a system where a human governor translates information for an artillery valve — resulting in a controlled, predictable shower of shrapnel and parachutes.

Does this creep you out? It should. In this scenario, the human actors are abstracted away. Not in an “assume the horse is a sphere” way, but in a “the human actors cancel out entirely” way. The pilot, constrained the physical capabilities of his craft and need to avoid passing out, can be reduced to Brownian probability, no more intelligent than an oxygen molecule. The AA gun operator is reduced to a pattern recognition machine. “Point at the plane, the math will do the rest”. Errors in the system are smoothed over with a suitably wide blast radius.

While Wiener was working on the AA gun problem, he wrote a 120 page report called The Extrapolation, Interpolation and Smoothing of Stationary Time Series. It was immediately classified and came to be known as the Wiener’s Yellow Peril. In it, Wiener establishes the “fundamental unity of all fields of communication”. Wiener doesn’t distinguish between mechanical, electrical, biological, or cultural communication.

The same mechanisms that regulate a steam engine and aim artillery can be used to manage politics, under this conception. The math has already been done, the only question that remains is what the constraints are on the populations cultural motion. Knowing what that is (or what that even means) allows one to build a predictor and a regulator that should give you a controlled, even output.

The (underrated) second Matrix movie paints an excellent picture of what that might look like. Do you remember the ending of the second movie? Neo meets the Architect (who is a fan of overly large words). The Architect explains that the Matrix needed a pressure-release valve for handling a certain minority of resistors. So they set up a colony of rebels in Zion. The rebels rescue more trouble-makers from the system acting as a kind of liver for the Matrix, cleansing it of their toxic presence. Every now and then the machines go and kill everyone (who are conveniently all in one place) and reset the system.

In one fell swoop, the second movie makes every moment of the drama in the first movie entirely irrelevant. The humans rebels — with all of their hopes, dreams, and triumphs — cancel out. The details of are unimportant, they are brownian motions in the larger regulated plan. This is the same political pessimism of Orwell’s Animal Farm. The pigs and humans are interchangeable parts.

The machine runs on.