Don't Mess With Me

Why fight when you can just look scary?

• Peter Nelson
• Jun 01, 1994

A mailman is delivering the mail. Let's say his name is Larry, and he's filling in for his friend Bob. He comes up the walk and sees on the porch a small dog eying him warily from the side and growling. He's seen dogs before, and this one doesn't seem terribly threatening, not barking outright, so he decides to go about his business. He smiles and says, "Nice doggie."

And the dog bites him.

"I don't get it," he later tells Bob.

"You did five things wrong," Bob says. "First, dogs don't speak English. They speak Dog. Second, you smiled and bared your teeth. To a dog, that's a threat. Third, you probably made your eyes wide and looked straight at it. That's also a threat. Fourth, you entered its territory without slowing down or turning sideways, which escalated the conflict. You're supposed to show some fear, or at least stop if you stare back. And fifth, nothing personal, Larry, but your ears stick out. You might as well have worn a sign saying 'BITE ME.' "

Larry is learning what Charles Darwin observed a century and a half ago-that it's a dog-eat-dog world. According to the laws of natural selection, creatures constantly compete, and only the fit survive. Yet to put it more accurately, researchers have been learning that the world is more dog-eat-rabbit (or whatever else the predator can catch) than dog-eat-dog. Within a species, canine in this case, the interaction is usually more like dog-warndog or dog-negotiate-with-dog, communication that ideally convinces opponents to back down or leave.

Behaviorists call such negotiations agonistic behavior, and they are finding that it exists in animals ranging from wolves to salamanders, that it often involves critical bluffing calculations-and that some of its true purposes may have little or nothing to do with threats at all.

The basic evolutionary function of threat displays is to maximize the benefits while minimizing the costs. "Whenever two animals come together and contest over resources, they show elements of both attacking and escaping," says George Barlow, longtime aggression researcher and biology professor emeritus at the University of California, Berkeley. "There's always a benefit, and there's always a cost," he says. The resources in contention range from food, to mating partners, to sites for raising young. If push comes to shove, an animal can break an antler, a tooth or much worse.

In some cases, the costs come simply from drawing notice. Take the signals from "a frog down in Panama that makes a chucking sound which frog-eating bats have learned to home in on," says Barlow. "A guppy in the streams of Trinidad turns bright orange, but that makes it easy for predators to find it."

For the most part, what threat displays say and how they're made depends on context. Think of the natural world as a big high-school cafeteria, but without the food waste. Animals display to identify themselves, establish dominance over a territory or protect access to a resource. A moose in Maine lowering its head and shaking its rack to drive off a rival is the equivalent of a senior telling a freshman, "Out of my seat, twerp, and don't stare at my girlfriend." In nature, as in high school, newcomers-which upset the social order-get the full treatment.

How animals display threats can be incredibly varied. Many are visual, which means body language. A threat can be made in a pose, in the distance between adversaries or just in a way of walking. If you're hiking in the Maine woods and notice a moose walking parallel to you, you might want to put on a fresh application of moose repellent; among many animals, taking up a parallel position to a rival is a way of challenging it.

"If you match two pike of similar size in a tank," says Berkeley's Barlow, who has most recently observed fish behaviors, "they'll line up parallel, puff out their throats, snap their jaws and beat their tails at one another." If you match pike of different sizes, "the larger one begins to turn toward the smaller one. If the size difference is great enough, the larger fish will grab the smaller fish and eat it." The pike puffs up to exaggerate its size, much the way birds ruffle feathers or canines raise fur on the backs of their necks.

An illusion of size is just one type of threat. In many species, from horses to bobcats, the ears rotate back to show increased aggression and flatten out just before an attack, tucked away from the slashing and tearing of tooth or claw.

Then there are eyes, which widen to expand the range of vision when a challenge is perceived, blink at stress and frown or glare to protect themselves when danger increases. Some agonistic conflicts become prolonged stare downs in which the first animal to look away is the loser. The fish Badis badis, in Burma and India, is a master of what Barlow calls "color fights," common in diurnal fishes. "They'll go totally motionless, about three or four body lengths apart, and start changing colors while staring at one another," says Barlow. "Finally one moves away."

Displaying animals often brandish their weapons. Canines pull their lips back in a defensive snarl or up in an attack mode, which was what Larry the mailman saw. Many animals have evolved conspicuous colorations to enhance their threats, such as a wolf's black lips, which make the teeth more prominent when bared, or the patterns around a tiger's eyes.

Animals also vocalize their threats. At sunrise on a spring morning, bird songs that may sound peaceful and relaxing to humans are in fact nasty remarks birds make to each other as they establish their nesting territories. A dog's low growl is partly an attempt to make itself sound bigger than it is; the lower the tone, the larger the instrument making it. During autumn rut, hull elk can roar 10 times a minute for an hour or more. When two hulls face off. each seems to be trying to get in the last word, and because roaring is physically exhausting, the fittest animal usually prevails.

Scent particles, or pheromones, also advertise a threat. Animals that mark with urine are. in a sense, using the written word. House mice mark almost constantly, with the subordinate males peeing everywhere, and the dominant male of a family group then counter-marking the subordinates' trail. The strength and ubiquity of the background aroma intimidate trespassers, as if the dominant male were saying, "Not only am I tough, but look at the size of my posse."

Aggression could well be universal. liven amphibians have turned out to be aggressive. In the 1960s, researchers "were startled to find," says Barlow, "that bullfrogs light, as do some tree frogs and salamanders: it was just a matter of people not having observed them enough."

Though aggression seems innate, its proper use can take some education. In a 1989 Mitch study of black-headed gulls, birds raised in isolation were as aggressive as birds raised in social groups. however, the birds raised in isolation lacked the display vocabulary needed to socialize properly, and they fought or showed fear inappropriately. Some birds raised in isolation even attack themselves, for lack of a partner, or chase their own moving tails.

Aggression is often cued to an external stimulus such as the moving tail. Richard Coss, a psychology professor at the University of California, Davis, has been studying California ground squirrels to determine just how early in an animal's development it is able to recognize danger or threats. This ability may also be innate, since the sooner a newborn animal can recognize a predator or read the intentions of another member of its own species, the better its chances for survival. In experiments, some newborn animals are fearful of specific stimuli. Monkey newborns react to snakes, for example, and baby rodents to cats.

"Humans recognize eye patterns at birth," says Coss. "Infants often find it uncomfortable to be stared at. They look away, or gaze-avert." Jewel fish, he has found, do the same, about 13 days after being spawned when their optic tectal neurons become adultlike. "They avoid the front of the parents' faces and school next to them," he says. "They're also frightened of models with two eye spots, but not of models with other combinations of spots." He adds, "Recognizing eyes is clearly built into the brain."

The point of making threats is to convey information about future behavior: "If you do that, I will do this." Early ethologists first assumed that when animals display, they tell the truth about their fighting abilities, and that display contests produce the same results as would actual combat. The whole idea of natural selection, after all, is for the strong to prevail and pass their genes along to future generations.

For decades, that idea seemed to hold up in studies of the layers of information contained in a show of strength. Oxford University professor Niko Tinbergen, one of the pioneers in the study of the functions of threats, is known for his work with gulls, published in the 1950s. Displays, Tinbergen held, carry two messages, the gull first demonstrating feats of strength or agility, stomping the ground or furiously pulling up blades of grass with its beak, as if to say, "This is what I could do to you and your feathers."

The second message says, "This is how much I feel like doing it," expressed by the form, frequency or duration of the display. Tinbergen also observed that many display contests contain "graded behavior," which entails a back-and-forth "I dare you/I double-dare you/I doubledog-dare you" ritual in which the ferocity escalates until the weaker individual surrenders or withdraws.

Traditional interpretations came into question in the early 1970s, when John Maynard Smith, an English mathematician, applied to animal behavior a mathematic model called game theory, used in economics to describe how parties with conflicting interests resolve difficulties. Smith concluded that animals must lie, since keeping an opponent guessing can be strategically wiser than telling the truth about what you're going to do. Among humans, one analogy is looking confident when holding a losing poker hand. If the biggest lie reaps the greatest reward, animals would evolve to tell the biggest lie all the time, consistently using the one most effective bluff behavior.

In theory, Smith's conclusions made sense. But in practice, many animals have a wide repertoire of displays, not the single display Smith said would evolve. "It really upset behaviorists," says Alan Bond, a research associate at the University of Nebraska State Museum. Soon the big question in the field, says Bond, was "why there were so many different kinds of displays."

To understand game theory, think of animals playing poker. Originally, behaviorists thought animals play five-card showdown, with the cards dealt face up and each animal betting greater amounts (graded behavior), hoping its opponent will fold before the final card-actual combat-is dealt. In Smith's theory, the game is more like seven-card stud, with some cards dealt face-down to hide assets or conceal handicaps so an opponent can't devise a counter strategy. An animal cannot afford to admit, "I'm too old for this," since telling the truth might get it killed. Better to lie.

The theory contained a paradox. If two gulls fighting over a patch of beach are both lying about how tough they are, then the biggest bluffer will win. But if bluffing wins contests, then bluffers will procreate, everybody will start bluffing and all bluffs will become meaningless or noninformational. And then the only actual way to determine dominance would be to fight, which is what the gulls are trying to avoid in the first place.

A bluff, therefore, has to have some relationship to ability and must meet what Bond calls the "optimal level of deceit." In human experience, we know that exaggeration only goes so far. Iran's late Ayatollah Khomeini once called U.S. President Jimmy Carter a "cadaver eater," when most people agree Carter would never knowingly eat a cadaver. If Iraq's leader Saddam Hussein had promised during the Gulf War to "swim in the blood of American school children," he might not have endeared himself to your PTA, but he wasn't going to scare any body into overreacting. If, on the other hand, he had said, "Tomorrow at noon, I nuke Tel Aviv," he might have drawn more of a response.

Among animals, "true deceit is rare," says Berkeley's Barlow. "At some point, the displayers get tested, and if they're deceitful, they get whipped. When fox sparrows molt, the dominant birds grow in a larger black bib on their throats. You can take subordinate birds and artificially make their bibs larger, but those birds really get beaten up."

Beyond the bluffing question, researchers are even questioning their assumptions about the reasons for displays. Bond, for one, believes communication may be only partly the purpose of threat displays. Indeed, gulls sometimes display without apparent stimuli, when no other gull is watching or even present. Why?

First, a threat display could be a kind of regulating mechanism, an animal's way of counting to 10 before doing something it might regret. "It's disadvantageous to fly off the handle," Bond says, "especially when, most of the time, the other individual is a mate, a relative or an offspring. Evolutionarily speaking, it's not a good idea to go around creaming your kids, though one occasionally has the urge to."

At the same time, display rituals may be how animals work themselves up to a fighting pitch, much the way football players pound benches in the locker room before a game. Bond calls that "behavioral efference," action that generates emotion, not the other way around. Behavior modification experiments with humans show we can control our feelings by manipulating our behavior; if we're unhappy, forcing our mouths to turn up at the corners can induce happy feelings. Elk bellowing at each other, once thought a way of getting aggression "out of their systems," may in fact be a way of getting it into their systems. So whether or not the animal is communicating may not matter, as long as it is getting ready to attack or defend.

Communication still remains a component, however; threat displays are usually two-way conversations in which animals both send and receive signals. There's a symmetry to such dialogs, aggressive signals met with submissive signals that follow what Darwin called the principle of antithesis. A common example: If a dominant individual holds its head high, the submissive one holds its head low.

Many displays have an internal symmetry as well, comprised of aborted or incomplete attack-and-escape sequences. Tinbergen found that gulls reflect their internal indecision to fight or flee by raising and lowering their heads and wings, formalizing the behavior until it becomes a kind of dance, a chain of signals indicating willingness to attack or fly away. The moose walking parallel to you in the woods is sending an equally indecisive symmetrical signal, neither advancing nor retreating.

For humans, one obvious reason to study how animals avoid conflict is perhaps to learn how we might do the same. People, after all, vie for limited resources like any other species. Barlow, for one, points out that a battleground is apparently developing over plain water, saying that during the next century, "you'll find the world fighting like crazy over water." Whether that prediction holds or not, if we could recognize the early signs of aggression and deter its escalation, the next time we catch ourselves bathing in the behavioral efference of a Cold War, we might stop before we get to where it takes 20 years just to disassemble the weapons we've built.

As for conflict at the level of Larry the mailman's daily routine, suppose he returns to the house where he was bitten. The same dog "grrrs" at him in the same way. This time Larry canidomorphizes and calculates, "This dog doesn't like surprises or the possibility of getting hurt any more than I do." Larry slows his approach, doesn't smile or widen his eyes, and he's wearing a hat over his ears. He pauses to allow the dog to de-escalate his own display. The dog sidles up to him, sniffs him to identify him ("Oh, this is the idiot I had to bite yesterday") and then the tail starts to wag. The point is, the second time around, Larry learned the dog's language, and in this case, it made all the difference.

Peter Nelson lives in Northampton, Massachusetts, with a dog so submissive that mailmen bite it.