The First Word: The Search for the Origins of Language

by Christine Kenneally

  The vervet story invokes many of our muddled ideas about animal communication and how it compares with human communication. A few themes crop up again and again. There is the notion that animal vocalizations are just gibberish, the opposite of language, and much like what we produce if we cry out nonverbally—informationless sound that provides a crude guide to an emotional state. And there is the contrary idea that animals use a code to communicate with one another, as we do, but we just haven’t cracked it yet. Both these approaches assume that animal communication will be recognizable in the terms we use to understand our own language, that it has words, or it doesn’t. It has syntax, or it doesn’t. It is full of meaning, or has no meaning and no reference whatsoever.

  Another suggestion is that other animals may communicate using degenerate, primitive tokens of our own language. This is part of the broad assumption that humans are intellectually and communicatively superior to all other animals. The vervet alarm calls seemed to fit nicely into this concept: the monkeys had words, but just three of them.

  It is true that the complexity of human language is without parallel. It enables us to connect with one another in a virtual world and together invent agriculture, construct buildings, send airplanes through the sound barrier, and shoot satellites into space. But assuming that the most salient thing about human language is that it is the superior form of animal communication doesn’t get you very far. It doesn’t tell you what parts of language may have been positively selected. And it can’t tell you about how language evolved. It implies that anything can be expressed by human language, when we don’t know if this is in fact the case.

  This approach also implies that human language is the communication tool par excellence, as opposed to a communication tool that developed in a certain niche. But assuming that language is the best possible communication tool is a little like saying that the human brain is quite simply superior to all other brains, as if our brain was an all-purpose machine rather than a device that does some jobs very well and others less so. Such evaluations take the trait, like language, or the organ, like the brain, out of the context of the body and the niche, as if evolution acted independently of the needs of the organism in its environment.

  Indeed, saying that the only important thing about language is what it does better than other communication systems is as nonspecific and unhelpful as saying that humans are the most intelligent species—which is itself like saying humans are the best-looking species. If you understand this sentence, then you already belong to the species that agrees with the sentiment.

  What matters about the alarm calls is surprisingly obvious but until now has rarely been commented on: when vervets and chickadees and chickens make their alarm calls, they are connecting a particular sound to a referent in the world. Whether the animal arrived at this behavior genetically, somewhat genetically, or not, it appears that it is a widespread, easily evolved, and useful trait.

  “Every species where researchers have tried—and that includes dogs, dolphins, parrots, and chimps—can link sound and a reference,” said Fitch. “I don’t think this is some sort of special human ability. It’s a pretty general ability. What else is your brain for? If your brain can’t link two stimuli in the world, one which is visual and the other which is auditory, then what good is it? I wouldn’t be surprised if fish could do this, but no one has really tried to see if they can.”

  So the act of hearing a particular sound and making meaning out of it is not particularly human; it’s ancient. Animals like vervets use the connection between a sound and a visual signal in one way, and humans have built on this ability in another way, using it as a platform for human language.

  In order to progress, science has to focus closely on some areas to the exclusion of others, which sometimes means that the most obvious facts of our daily lives are ignored. For instance, humans communicate with dogs. That observation is so mundane it hardly deserves mention, but this ability is relevant to understanding what evolved in order for us to evolve language and how the platform for understanding a word is ancient. Philip Lieberman spoke about the relevance of this ability in dogs in his 2000 book, Human Language and Our Reptilian Brain, but only in the last few years have other researchers begun to actively investigate it.

  In 2002 a team of researchers at the Max Planck Institute in Germany showed that Rico, a Border collie, knew the meaning of hundreds of words. Not only was Rico able to go into another room and retrieve an object he had been asked for (choosing it from a selection of possibilities); he was able to infer the meaning of words he’d never heard before. For example, Rico was asked by the researchers, who used a word he didn’t know, to go and retrieve an unfamiliar item. When he went into the next room to look for it, only one object in the set of possible things to retrieve was one he had never seen before. Because Rico knew the words for all the other objects, he picked up the novel one, assuming it was what the experimenters were asking for. Rico, obviously, does not have human language. Instead he is using an ancient, more general skill that preceded language by millions of years. It’s at this level that humans and dogs communicate with each other, as with an animal like Alex, the African gray parrot.

  Other animals appear to have built on the ability to make meaning from the connection between a sound and a referent, as with human words. Dolphins use echolocation clicks, “burst-pulse sounds,” and different types of whistles. “Signature whistles” are so named because it appears that dolphins name themselves.2 These beasts reproduce a distinct, individual sound that develops in their first year of life whenever they meet another dolphin. It’s always the same, and always distinct from any other dolphin’s whistle. There is even some evidence that dolphins will exchange their signature whistles when separating. In 2006 a team of researchers led by Vincent Janik at the University of St. Andrews in Scotland found that wild dolphins recognized that a signature whistle referred to a particular dolphin even when its voice was completely distorted.3

  Elephants also appear to use sounds like words. Katy Payne, lead researcher on the Elephant Listening Project at Cornell University’s Bioacoustics Research Program (now retired), and Joyce Poole, scientific director of the Amboseli Elephant Research Project in Kenya and another longtime elephant researcher, began an elephant dictionary study. The goal was to describe the way that individual elephants produce distinct sounds for various purposes, like greeting a fellow member of the clan they haven’t seen in a while. Dolphins and elephants don’t have words as we do, but both of these socially complex species have instead hit upon some of the same tactics to communicate.4

  Chimpanzee pant hoots are another interesting wordlike call. The pant hoots are very loud cries that are most often used to communicate over distances. Their function seems at least in part to be to rally support and keep individuals in a group together. Pant hoots also differ between individuals and between different chimpanzee groups. Chimpanzees appear to be able to pick and choose which ones to use. They are somewhat like dolphin signature whistles because they seem to have an internal structure and are uttered in various situations, such as resting, feeding, and during travel and display. This suggests that chimpanzees have some ability to choose meaning, as well as use structure.

  Klaus Zuberbühler and Katie Slocombe of the University of St. Andrews recently investigated the ability of chimpanzees to make humanlike reference in an experiment at Edinburgh Zoo. The researchers monitored the chimps and found that they issued distinctly different cries in response to finding different kinds of food. When the chimpanzees came across highly valued food, like bread, they made high-pitched grunts. When they came across food that was less appealing, like an apple, their grunts were low-pitched. Zuberbühler and Slocombe demonstrated not only that the chimpanzees were making distinctions in the way they vocalized about their food but that other chimpanzees seemed to understand the meaning of the different grunts. When they played recordings of the various food grunts to chimpanzees, the listeners would search f
or the given food in the place where it was usually found in their pen for a longer time and with more effort than in other spots where different food might be found. The chimpanzees also searched longer if the cry signaled a particularly prized piece of food.

  These findings suggest that our closest relatives have built upon the sound-referent connection to communicate distinctions to one another in a similar way that we do. There is more than just genetics involved with pant hoots, as well as signature whistles, and time and more research will help tell us the ways in which these sounds and meanings resemble human words. Certainly, it looks as if the voluntary production of sounds that are meaningful to another creature is not a uniquely human ability.

  Another question raised by the vervet studies is whether vervets intend in any conscious sense to communicate. Seyfarth and Cheney have carried out experiments on captive vervets in which they exposed adult females to a “predator” when they were either with a juvenile offspring or with an unrelated juvenile. The females gave many more alarm calls in the former case than in the latter. They also observed, in the field, one instance when an isolated vervet was being pursued in a tree by a leopard. The vervet gave no alarm calls, suggesting that the animals can withhold calls when no other monkey is around.

  As for the interpretation of the alarm cries: the monkeys learn what each cry means only through experience. (Some researchers argue that the cries induce only an emotional, not a cognitive, response in other monkeys, and that’s why they run. But as Seyfarth and Cheney point out, the fact that the responses are obviously emotional does not rule out the possibility that information in them is interpreted as well. Wouldn’t you also feel panic if someone screamed, “Snake!”?)

  The ability to interpret another’s animal utterance is so universal that even animals of different species can understand the cries that other creatures make. Seyfarth and Cheney have observed that predators that hear the alarm call of their prey often give up the hunt at the sound—they know they’ve been seen.

  Said Fitch, “What I think is interesting and surprising and we didn’t know twenty years ago is that animals have an asymmetry between perception and production. This appears to be one of the key differences in being able to communicate with words and not.”

  While other mammals appear to be very good at making meaning from sound-plus-reference combinations, they don’t necessarily produce new sounds in connection with new objects in the way we do. “The intuition is that if you can see something you must be able to produce a word for it,” Fitch explained, “but that’s where the data is completely clear—it’s not so. Dogs can bark, but they do not create new barks to correspond to new sounds, and chimps can scream, and they can even withhold their screams in certain contexts, but they can’t freely create new screams to correspond to new things.”

  Clearly there are varied ways that ancient capacities are used by different species, but being able to both understand and produce words is one of our special talents. Over time, we have produced hundreds of thousands of words, and there is little evidence that animals naturally produce many wordlike tokens at all. Individually we learn tens of thousands of words in a lifetime, and if we want, we can make up as many as we like. Language is in constant flux, so regardless of our own individual contribution to language change, words do inevitably become altered over hundreds of years. Rico and Alex and many other animals are able to comprehend that new sounds can refer to new objects, but they are not even remotely as adept at inventing words themselves.

  Human words are much more than just links between sound and reference in the world. Indeed, reference is not the half of it. A word is an arbitrary association between sound and meaning. There is nothing in the sound of a word that tells you what it means or what it does—you must learn this as a child.5 Whenever you hear a word, you know what (if anything) it refers to. You know that some words stand alone, like “hello,” “ouch,” and “yes,”6 and that others can join together to create larger words, like “heretofore” and “bedroom.”

  You know that a word is a noun or a verb or another part of speech. If it’s a noun, you know how to make it plural or singular. If it’s a preposition, you understand that it relates two nouns together in space or time. If it’s a verb, you know how to render it in a handful of different tenses, and you know what nouns will make sense with it and which ones will be nonsense. You know that some verbs have to have agents, like “killed.” All of this information about a word is specific to language. Of course, you know that “table” refers to a table, and on this level learning a word and learning an object may not be dissimilar processes. But all of the information about the way a word combines with others in language is internal to the language that you learn.

  You may not consider most of this consciously, but when you learned language, you internalized all of this information, and when you hear any word, you use this knowledge in the way you process it. You know all sorts of things about just the sound of the word. You know what other words will rhyme with it. You know which words start out with the same series of sounds, even though they end differently.

  A child’s ability to learn many words is so completely different from anything observed in other species that many researchers propose that some neural mechanism must be especially dedicated to this acquisition of linguistic knowledge.

  Beyond the basic link between an unanalyzed sound and a simple reference in the world, words are clusters of complex knowledge about sound, grammar, and meaning. Human words don’t exist by themselves. They are points in a series of intersecting systems, and when you hear or produce a word, all these systems come into play. Recent research has shown that when children acquire words, they are not just creating a multidimensional connection between different kinds of linguistic and nonlinguistic knowledge based on a platform of sound and meaning. The essential scaffold for word learning is more complicated than that. As well as a connection between two domains, such as the aural and the visual, there is a very important connection between speaking words and gesturing meaning.

  7. You have gestures

  Picture the house in which you grew up. Think about the rooms, the hallways, the stairs, visualize where they all are. Where was the front door? The back door? What color was the roof? Did you have wall-to-wall carpeting or were rugs spread all over the place? If you turned now and attempted to describe the house to someone nearby, it’s highly likely that you’d gesture as you spoke. In fact, even if you just imagine a person and then describe the house aloud to her, you’ll probably gesture as well. Gesture experts say that it is almost impossible to talk about space without gesturing. Gesture is spontaneous, and as integral to individual expression as it is to communication. Even though you probably won’t gesture as much if you are talking on the phone, you will still wave your arms about. Blind people gesture when they speak in the same way that seeing people do.

  Gesture may be integral to human expression, but it is not uniquely human. At the Gestural Communication in Nonhuman and Human Primates conference in 2004, Mike Tomasello of the Max Planck Institute in Leipzig, Germany, and his associates presented a huge compilation of gestures that they had observed in monkeys, gibbons, gorillas, chimpanzees, bonobos, and orangutans. Many of them had been observed at the spectacular ape exhibit at the Leipzig city zoo, where a leafy path leads to the center of a big ring. Radiating out from the central space are walks that divide all the great ape species from one another. In one section are the gorillas, sitting impassively. In another are the bonobos—only three of them, a reflection of their dwindling numbers worldwide. In the third section are the orangutans. The male sits near the viewing window looking profoundly deflated, while his orange cage mate hangs upside down from a tree stump and stretches. In the fourth section are the chimpanzees. There are more than a dozen chimps in the compound, and they make a lively community. Some recline sensuously, others fly through the air on ropes or trunks. Some busily work at boxes, inserting sticks into various holes. T
he exhibit is climate-controlled; it feels like a light summer day. Tomasello has a number of testing rooms installed at the zoo for his various experiments.

  Gestures play a large role in primate communication, Tomasello explained, and as is the case with humans, these gestures are learned, flexible, and under voluntary control. Most primates, humans included, gesture communicatively with their right hands, suggesting that the dominance of one side of the brain for vocal and gestural communication could be as old as thirty million years. Just as with human gestures, ape gestures can involve touch, noise, or vision. Apes wait until they have the attention of another ape before making visual gestures, and often if their visual or auditory gestures are unacknowledged, they will go over to the ape they want to communicate with and make some kind of touching gesture instead. Apes also repeat gestures that don’t get the desired response. Like human gestures, ape gestures seem to be holistic: a series of gestures doesn’t break down cleanly into meaningful components. Moreover, a set of different gestures may mean just one thing, while a single gesture may be used to convey many meanings.

  Tomasello and his group divide ape gestures into two types: attention getters and intention movements. Attention getters, said Tomasello, slapping the podium, do just what they say—they call attention to the ape making the gesture. Chimpanzees will hit the ground, clap their hands, and stamp their feet for this purpose. They also lay their arms on other chimps, tug on their hair, or poke them. Once the observer pays attention to the gesturing ape, said Tomasello, what is required becomes clear. To illustrate this, Tomasello showed a video of a chimpanzee who walks over to another chimp and starts jumping up and down on the spot. When the second chimp finally notices the display, the first one turns around and sits down. The message is obvious—groom me, and that’s what the second ape starts to do.