You May Also Like (26 page)

For a tangible human example of what was happening with the birdsong, consider irregular verbs in English. Why have some been converted, over time, into “regularized” verbs? And why have some stayed irregular? As the data scientists Erez Aiden and Jean-Baptiste Michel note, we no longer say that something “throve” but that it “thrived.” Using a database of English texts, they found that the more often an irregular verb was used, the more likely it was to stay irregular. Why? Because the irregular verbs we hardly ever encounter are the ones whose irregular forms we are least likely to remember; hence we convert them, through error, into regular verbs.

It is, they suggest, a process of cultural selection: “
The more frequent a verb is, the more fit it is to survive.” No conspiracy sought to kill off “throve,” nor is “thrived” any more inherently appealing. “
Thrived” thrived because people simply had trouble remembering the rarely used irregular form. People made mistakes, “thrived” got copied, more or less randomly, and, presto, over a few hundred years the past tense of “thrive” got changed into something new, much as the warblers' songs did.

This raises the question of how much our tastes evolve, on the wider social level, due to more or less accidental, random processes, cultural “mutations” that are not necessarily better, just different.
Music is filled with moments where mistakes became innovations (for example, the rise of hip-hop “scratching,” Cher's exaggerated use of Auto-Tune in “Believe”), innovations that ultimately shifted taste. The first use of guitar distortion on a record is, like many creation stories, a matter of historical debate. Some guitarist no doubt had a piece of equipment
that malfunctioned—or maybe he simply turned it up too loud—and found some pleasure in the resulting imperfection. Then someone else likes what he has heard and decides to imitate it, while putting his own gloss on it, pushing the effect further along.

And so in a couple of decades you have gone from the slight (though certainly edgy at the time) buzz in a forgotten, proto-rock song like Goree Carter's 1949 “Rock Awhile,” to the meatier growl of the Kinks' “You Really Got Me” (fashioned by Dave Davies's taking a razor blade to the amp), to the full-blown howl of Jimi Hendrix (now electronically engineered via a custom fuzz box and big Marshall amps). No guitarist really knew he would like it until it happened; otherwise, he would already have been playing that way. Even Pete Townshend's act of smashing his guitar began as a “
complete accident.” As Bourdieu once wrote, “
To discover something is to one's taste is to discover oneself, to discover what…[o]ne had to say and didn't know how to say, and, consequently, didn't know.”

Taste change is like Wall Street's “random walk,” or the idea that the past is a shaky guide to the future. We expect convulsive change on the pop charts, but think of something like the most common colors in home furnishings, the most popular dog breeds, or the top baby names. In any given year, there would be a certain order. But this would almost certainly have been different five years earlier, just as it is sure to be different five years down the road. Could this turnover be explained, even
predicted
? I do not mean in the sense of which breeds or names or colors would rise and which would fall (because, per Wall Street's “efficient market” hypothesis, if we knew what was going to be popular, it already should be). But could the
rate
of change be predicted? That is the promise of what has been called the “neutral model” of cultural change.

The idea comes from a theory in genetics, revolutionary when it was introduced in 1968, which “
predicted that the vast majority of evolutionary changes at the molecular level are caused not by selection but by random drift of selectively neutral mutants.” In other words, most changes in genes just
happened
. They were driven not by external, functional selection pressures (for example, some factor of the local environment) but on their own, as if guided by some internal logic, one whose probabilities could be estimated.

When applied to culture, the “neutral model” says that something
like a list of breed popularity will regularly shift. Some dogs will suddenly become popular—not because some breed is inherently better than another or the upper classes suddenly favor one over another. Rather, popularity shifts through “random copying,” or one person wanting a dog because she saw another person with one.
This was what R. Alexander Bentley, an anthropologist at England's University of Durham, and his co-researchers found after they sifted through many years of breed registration data. Statistically, the dog breed popularity index follows a power law: A dozen or so top dogs command a majority of the registrations in each year. But
what
those dogs are is subject to change, and that change seems entirely random. A dog can rise from obscurity to popularity with no dedicated promotional campaign behind it; similarly, it can fall from popularity with no apparent explanation.

It is not as if the top dogs became popular, for instance, because they were intrinsically
better
dogs.
A study that looked at positive breed characteristics (good behavior, breed life, fewer genetic disorders) and breed popularity found no link between the two. Sometimes, those bred to be least healthy rise most in popularity (call it “unnatural selection”). Humans often do not even seem to pick dogs that are functionally adaptive for
humans
.
Harold Herzog, a co-author of Bentley's and a professor of psychology at Western Carolina University, notes that Rottweilers surged from twenty-fifth place to the United States' most popular in a decade. What followed, Herzog notes, was a steep rise in the number of people killed by Rottweilers and then, not surprisingly, a sharp subsequent decline in Rottweiler registrations.

There are certainly cases of selective pressure on dog breed popularity. One of the strongest is movies for children: After Disney's
101 Dalmatians
and
The Shaggy Dog
, dalmatian and sheepdog registrations rose.
The bigger the box office, the bigger the breed boost (though, in certain cases, the tail might have been wagging the dog, because the breed was already on the rise, which might be why it was chosen for the film). Movie tie-in breed fads, however, notes Herzog, are “the exception, not the rule,” and they have been losing strength. After Taco Bell's famous Chihuahua ad campaign ran, he points out, Chihuahua registrations actually
plummeted;
what was at least initially good for Taco Bell sales was apparently bad for the breed. What about winning the Westminster Dog Show? This, after all, was said to explain the “
fabulous rise in poodle popularity” in the 1950s. If it did then, it apparently does not
anymore: Westminster winners do not seem to move the needle, breed-wise, in the years after their win.

As with Edwin Long's
Babylonian Market
, whatever breeds are currently top dogs—and however much we would like to think they are there because they are somehow best—the only thing that can be predicted of future taste is that it will change. Once, in a top-floor conference room in an art college in London, I witnessed a top secret annual meeting held by Pantone, the color company, surrounded by color experts. These are people who do not just see the color black, but can amiably chat about the “family of black.” Their goal was to try to forecast what colors would be big the next year. Like movie producers in search of an ideal dog—one starting to show up at the margins but not overexposed—the colorists were attuned to what was already gaining some steam or being employed in a new way (for example, “a good navy is going to fulfill the role that black used to play”). Having found the spark, the company's color “forecasters” piled on the fuel.

When, for example, the company predicted orange for the summer of 2011, I was later told by an executive at Firmenich, the flavor and fragrance company, “you can look at what's out in the marketplace—this red orange, or flame orange.” It lurked on the new Camaro, the Sony Vaio computer, Hugo Boss's new Orange line. “You're connecting dots here that are traceable,” the executive told me. Like surfers, the forecasters were catching a wave that had already begun, and as with the complicated physics that explain those “
rogue waves,” which just surge “out of nowhere,” orange, in all likelihood, was just coughed up from an ocean of color possibility. Like rogue waves, popularity tends to be nonlinear: Once it gets going, it gets bigger than you would have been able to predict from its initial condition (rogue waves “steal” energy from surrounding waves; popular dogs “steal” momentum from other dogs).

What makes the neutral model so compelling, suggests Bentley, is that it provides a way of thinking, at the wider “population level,” about why things like tastes just seem to come and go. Statistically, the rate of turnover, on quite distinct indices of popularity—ranging from the
Billboard
Hot 100 to baby names to which “keywords” appear in academic papers in a given year—seems to look the same, as if there was some natural law of churn.

With baby names, Bentley argues that even as the population of countries grows, new names are created (and others disappear), and specific names rise and fall in popularity, the overall statistical shape of name popularity changes little, because of the way people randomly copy names from each other. Remember that from its origin in genetics, the neutral model says that genes cannot be under selection. They cannot be chosen for an “adaptive” reason, where one is “intrinsically” better than another. Are baby names, as Bentley argues, really “value-neutral cultural traits chosen proportionally from the population of existing names, created by ‘mutation' and lost through sampling”?

Baby names have long fascinated taste researchers. As Stanley Lieberson, a sociologist at Harvard University, has pointed out, names, unlike many other fashions, are generally for life. No advertisers are cajoling you into a particular name, and they are “value-neutral,” in terms of actual money. “
It costs no more in dollars and cents to name a daughter
Lauren
or
Elizabeth
,” he writes, “than it does to name her
Crystal
or
Tammy
.” Names, notes Lieberson, were once largely bound up in tradition and social strictures; one took a family name or a name inspired by one's religion—sometimes to the point where the naming pool was beginning to get a bit small. In the genetic model, they were strongly selected, particularly for boys (in nineteenth-century England, for example, a consistent flurry of Williams and Johns and Henrys). But in the late nineteenth century, names, like so many aspects of culture, were becoming increasingly based on individual choice: “on whether parents like or dislike the name.”

Names went from tradition to fashion. And fashion, argues Lieberson, is driven by two large and distinct forces. The first are external factors, big societal ripple effects, like the way the name Jacqueline began to rise in the United States in 1961, thanks to the prominence of the famous First Lady. These large external correlations often do not work, however. A rise in biblical names, Lieberson has found, corresponded to a
decline
in church attendance; what's more, the least religious people were using the names.

More important, he suggests, are the “internal mechanisms” that drive taste changes even “in the absence of external shifts.” In what he calls the “ratchet effect,” some new, small taste change is introduced (like a simple change of letter in a name, from Jenny to Jenna). A mutation,
as the genetic model would have it. Another taste change subtly expands on that, typically in a similar direction. So skirt lengths or hair get a bit longer and then a bit longer, until reaching some point of disutility—or just ridiculousness. It echoes Loewy: most advanced, yet acceptable.

The attempt may be made to pin a taste change, after the fact, on some social factor (X became popular because Y). But it is often hard to escape the sense of sheer randomness and copying. Lieberson notes that boys' names ending in
n
became popular in the second half of the twentieth century (possibly hitting their zenith in 1975, when Jason, dramatically less popular decades earlier, hit number 2). They then began to decline. What happened? It is not as if an
n
sound confers any more intrinsic worth or that we are biologically programmed to prefer
n
names—for why would its ascent have stopped? Rather, it is as if people, like those warblers, were hearing a sound, presumably liked the way it sounded, and so took it on themselves.
One statistical analysis, of a century's worth of names, found that names, even after taking their past popularity into account, were more likely to be used when a sound contained in the name was popular the previous year.

Once a sound is introduced to the naming picture, it opens the doors to “errors,” imitations that are slightly off: The popular 1970s moniker Jennifer, writes Lieberson, “generates interest” in a number of similar-sounding names (like Jessica). The event that kick-starts a popular sound can literally be a matter of chance:
A study of naming patterns in the wake of hurricanes—whose names are randomly drawn from a list—found an increase in names sharing the first letter of the named hurricane. The bigger the hurricane, the bigger the increase (up to a point), simply because the “phoneme” was thrust in the air.
This is not so different from the way a “genre” book that hits the
New York Times
best-seller list can boost sales for non-best-selling books of the same genre, as if once people had read one, they were subtly influenced to read others.

You are probably protesting by now that this makes it sound as if we are all mindless drones marching in lockstep, naming our kids based on something we overheard at the grocery store or on the Weather Channel, doing things without any conscious thought. Indeed, critics of the neutral model insist there is almost
always
some kind of biased
selection going on. Most common is popularity itself; what is popular gets reproduced because it is popular.