Should people or machines count votes?

This Tech Talk column originally appeared in The Exponent (University of Alabama in Huntsville), Vol. 32, No. 13 (November 30, 2000). Digitized issue: UAH LOUIS archive. Reproduced here courtesy of that archive.

“Should people or machines count votes?” is the central question in the recent Fight for Florida. The man-versus-machine question arises in widely disparate areas like manufacture, repair, education, and driving. Some even investigate the pos sibility of computers programming themselves (perhaps at the direction of a human). We’re glad to foist our menial and repetitive tasks onto ma chines, but how far are we willing to go in trusting machines? Secretary Baker said that machines are “neither Democrat nor Republican” and, hence, the best vote counters. If machines are fair and impartial, then should we replace judges with machines, too?

Rewind about eighty years, back when electronic computing was in its infancy. Scientists were beginning to realize the power inside the new curiosity. Man, the bloodthirsty ape, invents some thing and then figures out how to use it to kill more effectively. An early application for computers was to speed up the process of grinding through ballistics calculations. If you’ve taken physics, then you feel the pain of the poor army grunts who knew a little calculus: a projectile launched with velocity v at angle theta from elevation h… Now consider that real ballistics problems are much nastier than their simplified cousins in physics textbooks; one has to account for air resistance, windage, and so on.

Mathematicians were looking much further. Computers were found to produce repeatable and reliable results, and mathematicians wanted to exploit those traits to construct a rock-solid mathematics from a foundation of axioms, similar to the attempt of Descartes to construct a rock-solid solid belief system atop “I think, therefore I am.” Their idea was to build a computer capable of verifying and rejecting proofs.

The hunters pursued their elusive fox until Kurt Goedel sounded the horn with his famous proof of his incompleteness theorem. Goedel demonstrated that within any consistent formal system there will exist a proposition that can be neither proven nor disproven. A common example is the Liar’s Paradox. Consider, for instance, the following statement: this statement is a lie. If it’s true, then it’s false. If it’s false, then it’s true. (Feel free to take a break until your head stops spinning.) In effect, Goedel was the Alan Greenspan of the day, bursting the nice bubble that his friends were enjoying.

Pop back to the present and the question of whether machines should replace human judges. Goedel’s proof offered wonderful job security to judges and lawyers, but anyone who’s ever tried to resolve the inconsistencies (e.g., seen a deduction declared legal in one place and illegal in another) in the tax code knows that IRS agents and tax accountants may be in trouble.

Incidentally, many considered Goedel’s proof to be a crisis much like the situation in Florida is seen as a crisis. His proof didn’t trouble Goedel, however; instead it validated his belief that human intuition plays a vital role in the decision-making process.