Liars and Outliers (13 page)

Take appearance, for example. Numerous experiments indicate that we are more likely to trust
people who look like
us. The phenomenon goes well beyond race; experimenters have digitally manipulated images of faces to more or less resemble those of their subjects and found that a variety of prosocial behaviors are correlated with facial similarity.

Dialect is a particularly interesting marker of group membership. With the nationalization and globalization of mass media, both accents and dialects are fading, but for most of human history, they were localized.
12
They're hard to fake, unless you're a rare gifted mimic, and they're generally set by adolescence. There is a lot of evidence, worldwide, that people are predisposed to cooperate with someone who speaks the same dialect they do. For instance, in one experiment, subjects were more likely to trust people with the
same accent
they had.
13
And we naturally
change our patterns
of speech and body language to mimic those around us, unconsciously trying to fit into the group.
14
Of course, the flip side of this is that we're less likely to trust people who
don't sound like
we do. Again, dialect preference seems to be a vestigial kin-recognition system.

It's worth noting that membership markers are harder to acquire and fake for groups that involve long-term—even inter-generational—cooperation and trust, than they are for groups involving more near-term cooperation and trust. It's much easier to learn the knowledge and skills to be a member of the community of football fans, or stamp collectors, or a particular church, than it is to acquire a new facial feature like an epicanthic fold or a dialect.

Gauging reputation by group membership is a lousy way to prejudge someone—another name for the practice is “stereotyping.” But it's not an unreasonable cognitive shortcut, given our inability to interact meaningfully with more than 150 people, or even to put names to more than 1,500 faces. Historically, as the number of people we interacted with grew, we had to develop these shortcuts. Identifying someone as a member of a particular community, whether an ethnic community or a community of choice such as a professional association, gives us some indication about whether she is likely to cooperate with us or defect. If she's a member of the same community as us, we know she's likely to share the same set of ethical rules we do.

Recall the Golden Rule. It's not enough to want to cooperate. You also need to know
how to cooperate
according to your society's particular definition, so others can know you're reliably cooperative. One popular business-success book tried to “improve” on the Golden Rule, creating what it called the
Platinum Rule
: do unto others as they would want you to do unto them. That sounds even more altruistic, but it's not what has been encoded in our brains. Figuring out what someone else wants is easy to get wrong. It's much easier to assume that another person wants what you want. Of course, that works best if you only deal with people who are like you, and are likely to want the same things you want.

Social norms tell us how to cooperate. This is one of the reasons societies have tended to be homogeneous in their morals: it's advantageous. When people with different morals interact, they may have different default assumptions about what it means to cooperate. Remember the Machiguenga tribesmen in Chapter 7? They use a different definition of “fair” than Westerners do. Cooperation works better if we all agree on what it means to cooperate.
15

Of course, just because our brains are hard-wired for this sort of in-group/out-group division doesn't mean it's the right thing to do.
16
There are all sorts of reasons why stereotyping is a bad system for judging individual people, and for those reasons we should strive to get beyond our more base instincts.

A substitute that can help
reputation scale
is commitment. By committing ourselves to an action in a way that we cannot undo, we can make up for a lack of reputation. Consider the coordination problem between a prostitute and a prospective client. The two have met in a bar and have agreed to meet upstairs in his hotel room later in exchange for $100. She wants him to pay her in advance, because she doesn't trust that he will pay her in his hotel room as promised. Similarly, he is concerned that, once having received the money, she won't follow through and meet him later. If the two could trust each other, this would be easy to solve. But they don't.

One solution is to tear the $100 bill in half, one piece for each of them. In the U.S. at least, half a $100 bill has no value, so neither party has the money. Now both parties have effectively committed to the rendezvous: so she can receive the other half of the $100 bill and he can receive the service. If either one of them defects and misses the meeting, neither gets the money.
17
eBay escrow services serve the same function; they facilitate trust by forcing the buyer and seller into a commitment they can't get out of easily.

A similar mechanism is to deliberately cut off your escape routes, so you have no choice but to follow through on your commitment. This could mean literally burning your bridges behind you. In 1519, when Hernán Cortés invaded what today is Veracruz, Mexico, he scuttled the ships he arrived on, signaling to both the Aztecs waiting for him and his own men that there would be no reneging on his commitment.

A second way to demonstrate commitment is to move in steps. When I hired a contractor to perform renovation work on my home, the contract stipulated several partial payments at different milestones during the project. This step-by-step approach—me paying the contractor partially, him doing some of the work, me paying some more, him doing some more work, etc.—helped both of us trust each other during the entire project because the severity of defection was lessened.

This was also the general idea behind the Cold War doctrine of mutually assured destruction. Both the U.S. and the USSR worked to convince the other that they were committed to massive retaliation in the event of a first strike. The result was that neither side was willing to use nuclear weapons; the two countries might not have trusted each other in general, but they both trusted that the other side was crazy enough to follow through on its commitment.

A third way to signal commitment is ritual. This could be a handshake to seal a commercial deal, a ceremony to seal a marriage, or an Eagle Scout induction ceremony. Rituals work because 1) reputation is at stake, and 2) society provides sanctions against anyone who reneges. Of course, these only work if everyone understands what the ritual is and what it means.

Zahavi's handicap signals from Chapter 3 are another way to scale reputation: costly and hard-to-fake demonstrations of our reputation. These include publicly attending religious services to demonstrate our morality,
18
ostentatiously spending money to demonstrate our social class, and engaging in particular activities to demonstrate our political or cultural proclivities.
19
Nobilities have complex displays of etiquette. Banks spend some of their money on imposing buildings to show off their financial health.
Criminals have signals
, too, to advertise their “good” reputation as a career criminal: prison time (that fellow criminals vouch for), tattoos, and deliberate physical self-harm.

Branding is yet another way to make reputation scale, similar to group membership. In many cases, we interact with organizations as groups rather than as individuals. That is, the corporate reputation of McDonald's is more important to our decision about whether or not to trust it than the individual reputation of any of the stores or the individual employees.

Branding isn't necessarily about quality; it's about sameness. Chain restaurants don't necessarily promise the best food, they promise consistency in all of their restaurants. So when you sit down at a McDonald's or a Cheesecake Factory, you know what you're going to get and how much you're going to pay for it. Their reputation reduces uncertainty.

Advertising can be about persuading consumers to associate a certain brand with a certain reputation.
Shared brand names
serve as means of aggregating individual reputations into an overarching group reputation, which—if it's maintained in good standing—benefits all members of the coalition. Companies call attention to their age, their size, and the quality of their products and services, all in an effort to enhance their reputation. Witness the ubiquity of advertising boasting about firms' positions on environmental and workplace issues or contributions to worthy causes. Or the effort the principals of the
Saudi Binladin Group
construction company have spent trying to differentiate themselves from their terrorist relative.

In ascertaining quality, consumers will often rely on the cognitive shortcut provided by a brand name, and will even
pay a premium
for products with brand names they associate with a reputation for quality. One study of Bordeaux wines found that customers will pay a premium for bottles from a
more reputable producer
, even if the wine is no better. Notions of branding have leaked into individual reputation as well. Career counselors now advise professionals to “
cultivate their brand
.”

A final way to make reputation scale is to systemize it, so that instead of having to trust a person or company, we can trust the system. A professional police force and judiciary means that you don't have to trust individual policemen, you can trust the criminal justice system. A credit bureau means that lenders don't have to decide whether or not to trust individual borrowers, they can trust the credit rating system. A credit card relieves merchants from having to figure out whether a particular customer is able to pay later; the system does that work for them. Dunbar's number tells us there is a limit to the number of individuals we can know well enough to decide whether or not to trust; a single trust decision about a system can serve as a proxy for millions of individual trust decisions.

We have a lot of experience with this kind of thing online: ratings of sellers on eBay, reviews of restaurants on sites like Yelp, reviews of contractors on sites like Angie's List, reviews of doctors, accountants, travel agencies…pretty much everything you can think of. Social networking sites systemize reputation, showing us whom we might want to trust because we have friends in common.

This is an enormous development in societal pressure, one that has allowed society to scale globally. It used to be that companies could ignore the complaints of a smallish portion of their customers, because their advertising outweighed the word-of-mouth reputational harm. But on the Internet, this isn't necessarily true.
A small complaint
that goes viral can have an enormous effect on a company's reputation.

On the other hand, while these reputational systems have been an enormous success, they have brought with them a new type of trust failure. Because potential defectors can now attack the reputational systems, they have to be secured. We'll talk about this in Chapter 10.

Reputation
isn't an effective
societal pressure system unless it has consequences, and we both reward cooperators and punish defectors.

We reward cooperators all the time incidentally through our actions. We choose to do business with merchants who have proven to be trustworthy. We spend time with people who have demonstrated that they're trustworthy. We try to hire employees who have good reputations, and we promote and give bonuses to employees who cooperate. From a security perspective, friendships are mutual reward systems for cooperating.

The common thread in all of these rewards is participation. Humans are a social species, and we reward by allowing others to participate in the group: whatever it is doing, whatever benefits it is accruing, whatever status and credibility it has achieved. Our brains are hard-wired to need to participate; we crave the approval of the group.

We also punish defectors. And if participation is the canonical reward, exclusion is the corresponding punishment. In our evolutionary past, the most severe punishment was banishment from the group. As interdependent as humans were, this punishment was tantamount to death.

We still banish people today. We tell them we're no longer their friends and that they shouldn't come around anymore. We cut all ties with certain relatives, kick trolls out of online communities, and unfriend people on Facebook. On a different scale, someone with a destroyed credit rating is pretty much banished from the lending community.

Other punishments are less severe: physical violence, property damage, and so on. Sometimes we call this sort of thing “revenge.” Here's how
Maine lobstermen
deal with one of their group violating traditional territories:

Ordinarily, repeated violation of territorial boundaries will lead to destruction of the offender's gear. It is usual for one man operating completely on his own to first warn an interloper. In some places this is done by tying two half hitches around the spindle of the offending buoys; in other places by damaging the traps slightly. At this point, most intruders will move their traps. If they are not moved, they will be “cut off.” This means cutting off the buoy and warp line from the trap, which then sinks to the bottom where the owner has no chance of finding it…. A man who violates a boundary is ordinarily never verbally confronted with the fact of his intrusion. And the man who destroys his gear will traditionally never admit to it.
20