They've got your number


On this date in 1958, the first computer-arranged marriage took place on Art Linkletter’s show. The historic pairing began two years earlier on Linkletter’s popular show, “People Are Funny”. As Time Magazine reported, “Remington Rand’s Univac No. 21 turned Cupid, brought together a flesh-and-blood couple as scientifically selected ‘ideal marriage mates.’” Time also noted that the love match was based on a 32-item questionnaire developed by “The Father of American Marriage Counseling,” Paul Popenoe. The happy couple became engaged, and Art Linkletter offered to pay the airfare for their Paris honeymoon.

In an era when sites like eHarmony and Match.com run ubiquitous advertisements for their “scientific” compatibility matching, it’s easy to forget the long history behind this kind of analysis. Some researchers argue that a line can be drawn back as far as St. Valentine’s Day, which has its origins in the festivals of ancient Rome.

Ancient Romans celebrated an early spring fertility festival, Lupercalia, in mid-February each year. Young women placed their names in an urn from which bachelors would select the year’s companion. Often these pairings resulted in marriage. Later, in 498 A.D., Pope Gelasius declared Feb. 14 as St. Valentine’s Day, and the Roman lottery system – disparaged as an un-Christian practice – was banned.

While this historical provenance places modern computer matching as little more than old wine in a new bottles, there is a marked difference between simple lotteries and the “Univac” solution. The German philosopher, Ulrich Beck, summarizes the change in his book, Risk Society: Toward a New Modernity. Beck argues that modern society has shifted away from an economic and social paradigm based in the production and distribution of goods to one that emphasizes the technical and scientific distribution of risk.

In short, modern life in the Risk Society is reducible to one great big actuarial table.

Students take the SAT to demonstrate that they present an acceptable risk to colleges — their scores are held to be predictive of academic success. Banks look at an individual’s credit scores to appraise their risk in granting the individual credit. Even the evening weather forecast is an exercise in probability distributions — how likely is it that rain will fall here?

We also see this in much more subtle ways. Anyone who has ever purchased something from Amazon.com is doubtless familiar with the phrase “customers who viewed (what you just viewed) also viewed … In other words, Amazon is using a predictive algorithm couched in your browsing history to suggest additional items for your consideration.

This is the Risk Society writ large.

Of course, here too we’re building on very old ideas. In his ninth century work, Kitab surat al-ard (The image of the earth), the Persian mathematician, Abu Abdullah Muhammad, describes the Indian system of numbers and measurements. When the book was translated into Latin during the 12th century it was renamed “Algoritmi de numero Indorum.” Because the name was transliterated as “Algoritmi” he is indirectly responsible for the term algorithm.

Simply put, algorithms are systems of equations for solving problems — predicting risk, for example. Most often algorithms involve the use of computers. This is all well and good, save for the fact that the same tool used for predicting whether you prefer red roses to yellow can also be used to predict the likelihood that you’ll engage in unsafe or criminal behavior; or that you have a predisposition to certain types of illness; or a proclivity toward certain work habits.

This then is the real danger of the Risk Society. As the algorithms get better at predicting us, a kind of solipsistic over-confidence emerges. The probabilities become the reality. The totality of our worth is distilled to a number.

If we’re just talking about buying flowers maybe that’s acceptable, but for any of life’s more consequential decisions we may not want to be drawn so tightly.