When I was a young lawyer at my first big Seattle law firm,
my colleagues had a Roger drinking game. The challenge was to identify a
conversation topic that would reduce me to silence. I don’t think anyone ever
won – I was willing to offer emphatic opinions on virtually any subject, from
quantum physics to rugby, regardless of my lack of knowledge.
My foolhardy confidence can be attributed to a couple of
things. First, I was finally drinking alcohol after twenty-five years of Mormon
prohibition. And second, I had just
graduated from Yale Law School.
At normal law schools, where teachers assign real grades, the students
attempt to master various substantive areas of the law. In contrast, at Yale our
professors taught us to derive everything in the universe from a few first
principles, using pure reason. Those Yalies who deign to become lawyers learn “the
law” in expensive cram courses shortly before the bar exam. (University of
Chicago grads also try to use reason to derive everything, but from the wrong
principles.)
With deductive reasoning, you reach a specific
conclusion by applying the fundamental rules of logic to an inherently limited
amount of information. In the history of ideas, philosophers who prioritize
reason are called “Rationalists.” Here is an example of deduction: “I drive a
battered Kia minivan” + “I went to Yale” + “Only soccer moms drive a minivan for
149,737 miles, long after all three kids outgrew car seats” ⇒ “I am a Jewish soccer
mom.”
In contrast, with inductive reasoning you hypothesize
general principles based on specific observations. The focus is on the evidence
you gather, not the reasoning process. Philosophers who prioritize the role of
evidence are called “Empiricists.” Induction allows you to make predictions with
varying degrees of probability when applied to particular situations: “42 of
the 57 nebbishy guys in the YLS Class of 1990 are Jewish” ⇒ “There’s a 73.7 percent chance
my last name is spelled ‘Leischman,’ not ‘Leishman.’” [Ed.
note: the Leishmans, i.e. Clan McLeish, actually emigrated from Scotland to Utah in the 1850s,
not from Ukraine to New York in the 1890s.]
No one ever has perfect information. Human brains evolved to combine a spiffy new rational frontal
cortex with the older animal limbic brain and the ancient reptilian lobes beneath. We are wired to make decisions and/or initiate
automatic reactions without waiting for all the answers.
Management gurus like Amazon founder Jeff Bezos recommend
making 100% of your decisions with no more than 70% of the relevant
information. More cautious leaders, such as Colin Powell, agree with the 70%
ceiling, but suggest you should defer decisions until you have acquired a floor
of at least 40% of the relevant information. (I assume he was talking about his
first war with Iraq, not the second one.) General Powell never taught at Yale
Law School.
Numerous human interactions require spontaneous responses – from
witty cocktail chatter and oral argument on appeal, to comic improv and staying
in the closet. Deduction is a much sexier parlor trick than induction. Sir
Arthur Conan Doyle never described Sherlock Holmes patiently waiting at 221
Baker Street for the editor of some peer-reviewed medical journal to confirm Dr. Watson’s blood-splatter data set was sufficiently large to
justify a two and a half percent margin of error. Instead, employing deduction,
Holmes tells the Inspector to arrest someone familiar to members of the
victim’s household because of “the curious incident of the dog in the
night-time.”
Inspector: "The dog did nothing in the night-time."
Holmes: "That was the curious incident."
My first memory of Sherlockian deduction comes from a childhood
visit to my great grandmother in Salt Lake – my mother’s mother’s mother. I had
been sent to fetch some basic household object, like the salt and pepper
shakers. I looked around the strange kitchen, closed my eyes, visualized where
my mom would have put them if this were our house, and opened the correct
cupboard.
As with most gymnastics exercises, deduction is more fun
with an audience. At my first Seattle
Men’s Chorus rehearsal seventeen years ago, the Second Tenor seated
next to me told me his name was Todd Feldman. Recognizing that ol’ MOT look in
his eye, I introduced myself with “Hi, I’m Roger Leishman – but I’m not
Jewish.” Todd almost fell off his chair, amazed by my mind reading. [Ed. note: “MOT” means “Member Of the
Tribe.”]
Mental McGyvery is more impressive the less
information you have to work with. If you’re too risk averse to try deduction
in public yourself, you should rely instead on ostentatiously researched inductive
reasoning. It increases your accuracy, while giving you a face-saving out when you
turn out to be flat wrong. (Plus you can
always pull a Nate Silver and gloat over your own near-non-inaccuracy
compared to all the other even-more-wrong people.)
Masters of sophistry effortlessly pivot from deduction to
induction without anyone noticing they might have been wrong about something.
Grand Masters never need to pivot, because they are never wrong.
In recent years, I have developed an unYalielike interest in
data. It’s probably a symptom of logomania. Or the proliferation of top
ten lists and Buzzfeed surveys online. In general, more data leads to superior
reasoning, and more accurate conclusions. So I've amassed a couple of bookshelves filled with Freakonomics, probability treatises, and Malcolm Gladwell-esque pop psychology books.
As logomaniacs compulsively look for patterns,
part of our challenge is to make sense of the world in light of our own observations
and experiences. For example, since 1990 I’ve checked for impairments in my
judgment by regularly making a mental list of the hottest guys around me. I'm still a Sherlockian Yalie. But the era of Big Data has astronomically magnified the kalaidescope I see the world through.
It’s not just me. Data
is trendy, from statistically-based websites like Nate Silver’s 538, to books
and movies like Moneyball, which tells how data-driven analytics have changed professional sports. So there’s the obligatory counter
trend of articles
about the potential dangers from fetishizing data.
Fortunately, I’ve always been a pragmatic skeptic. I try not to fetishize either reason or
data. To the contrary, most of the volumes on my behavioral economics bookshelf are about the limits of data, reason, or both. My favorites tend to be when the data reveals a different answer than
the one “reason” would have suggested. Over time, the best cocktail chatter focuses
on some strange wrinkle in the pattern.
