A duprass, Bokonon tells us, is a valuable instrument for gaining and developing, in the privacy of an interminable love affair, insights that are queer but true. The Mintons’ cunning exploration of indexes was surely a case in point. A duprass, Bokonon tells us, is also a sweetly conceited establishment. The Mintons’ establishment was no exception.
Some time later, Ambassador Minton and I met in the aisle of the airplane, away from his wife, and he showed that it was important to him that I respect what his wife could find out from indexes.
“You know why Castle will never marry the girl, even though he loves her, even though she loves him, even though they grew up together?” he whispered.
“No, sir, I don’t.”
“Because he’s a homosexual,” whispered Minton. “She can tell that from an index, too.”
If you want to figure out if your first date will sleep with you tonight, what would you ask them? What would be the single, best, statistically significant question you could ask to determine his or her licentiousness? The web’s got the answer for you:
“Do you like the taste of beer?”
A positive answer to this question indicated a 60% greater likelihood to “be okay with sleeping someone they’ve just met,” on OKCupid.com, one of the largest dating sites on the web. Among the 50,000 other questions reviewed, this one’s promiscuity predictive power is unrivaled.
In the IIII hierarchy, this is an inference, and a remarkably important one for all of us — not just the two hormonally-inflicted junior brand managers wasting their afternoons away on the opposite ends of a dating site — because of what it means for what we can know about people and how we can know it.
In the novel Cat’s Cradle, quoted above, a duprass is “a valuable instrument for gaining and developing… insights that are queer but true. “
In Kurt Vonnegut’s telling, the beauty of the duprass is their spontaneous generation of striking insights – unexpected glimpses through an open window into the human being.
Adapting the word to express its output, let’s term a duprass any insight that is “queer but true”. While Vonnegut was able to write his duprasses into his novel, we’ll have to understand how the web manufactures them.
I suppose that if you seek “queer but true”, you could reverse engineer it such that you find all things true first, then discard the obvious, rote, standard, expected, boring or tiresome observations to remain with the queer.
Queer, in this case, must mean that correlation is not enough; we’ll need to find those truths that surprise, tease and titillate, those that are not ready accessible via logic.
For a start, let’s consider the production of answers.
Me personally, I’ve always liked surveys and have filled them out whenever I could. Receiving them in the mail, submitting them after a hotel stay, pleasantly scratching out the bubbles on a guest satisfaction quiz, tickled me. I remember the little thrill that would go through my bones when I’d read through the disqualifier notices – you must not be an employee of the client or the agency or work in any marketing or advertising company – and conclude for the hundredth time in a row that they didn’t apply to me – I was eligible to participate!
Prior to the Web, surveys were conducted differently. I remember in 1994 being called to do a 1-hour interview about my subscription to Fortune Magazine, and meeting with the researcher at the McDonald’s in Sorrento Valley, San Diego.
I enjoyed it immensely, though I was a little let down at the payoff : Do I get to see the results? May I weigh in on the conclusions? Can I help explicate the answers, Dear Fortune? All for naught.
The cost, frequency and serendipity of surveying was low; a manual process, without cross referencing, conducted in person, could only be so productive. And perhaps not all Americans were as eager to participate as I.
So what happens when the average person increases their output of recorded answers from one per year…
From the single starling, you can’t predict the beauty of the flock, suddenly appearing and flashing shapes and colorations…
…the queerly mesmerizing truth of the flock in flight; its duprass. The emergent behavior of the internet is to reveal us to be fascinating.
Hunch.com, the pioneer in the inferences business the way Google was the pioneer in the intentions business, packages duprasses for us. What can we surmise from these Hunch correlations?:
South Americans are more likely than North Americans to dream of being underwater or being able to fly; North Americans are more likely than South Americans to dream of being chased or attacked.
Cookie Monster fans believe in life after death and consider themselves “foodies”.
Bacon double cheeseburgers lovers are more likely than haters to be able to do 10 pull-ups.
And before I respond “tsk tsk” to your cry that “these are just silly games; parlor tricks for people who still have parlors”, may I point you to that section of “Innovator’s Dilemma” which presents prior exemplars of market entrants becoming market revolutionaries through low-utility innovation?
I was a political science undergrad, and I’d been fascinated by Thomas Sowell’s A Conflict of Visions, a brilliant tour de force that uses inductive reasoning to draw the lines between the political ideologies and show how their many different answers stem from the same underlying differences in conception of human nature and possibility. It is a supreme example of what the human Expert can do: forty years of study from a leading political philosopher results in a book of penetrating insights.
Compare this with the work of a cheeky dating site run by flippant 20-somethings, who have discovered that the question most likely to indicate whether you are conservative or liberal is:
Do you prefer the people in your life to be simple or complex?
Or, for good measure, who determined that answering the question “Is my date religious?” could best be achieved by asking the question:
Do spelling and grammar mistakes annoy you?
Far from being parlor tricks, the portent of these inferences is the destruction of the social sciences – the belief that we can understand human behavior and explain it through our rational faculties — and the rise of something far more powerful, fast, and compelling: the Inference Machine.
If a taste for hops predicts hopping in the sack, if grammatalogical correctness tops theology in predicting religiosity, if a simple question tells us more than four decades of complex study, we face an epistemological paradigm shift in our understanding of humans in the 21st century.
Each survey, each poll, each choice online creates more answers. When we have thousands, when we have millions, when we have billions of answers to these questions each day, what will they tell us about ourselves? And what questions will the answers give us?
And who knows? When we’ve answered the questions, connected the dots, and drawn the lines between faith and fact…