Why Chimpanzees Can't Learn Language and Only Humans Can

Herbert S. Terrace

• Publisher: Columbia University Press
• Year: 2019
• ISBN:
  9780231550017
• Topic: Languages & Linguistics
• Subtopic: Syntax in Linguistics
• Index: Languages & Linguistics

In the 1970s, the behavioral psychologist Herbert S. Terrace led a remarkable experiment to see if a chimpanzee could be taught to use language. A young ape, named "Nim Chimpsky" in a nod to the linguist whose theories Terrace challenged, was raised by a family in New York and instructed in American Sign Language. Initially, Terrace thought that Nim could create sentences but later discovered that Nim's teachers inadvertently cued his signing. Terrace concluded that Project Nim failed—not because Nim couldn't create sentences but because he couldn't even learn words. Language is a uniquely human quality, and attempting to find it in animals is wishful thinking at best. The failure of Project Nim meant we were no closer to understanding where language comes from.

In this book, Terrace revisits Project Nim to offer a novel view of the origins of human language. In contrast to both Noam Chomsky and his critics, Terrace contends that words, as much as grammar, are the cornerstones of language. Retracing human evolution and developmental psychology, he shows that nonverbal interaction is the foundation of infant language acquisition, leading up to a child's first words. By placing words and conversation before grammar, we can, for the first time, account for the evolutionary basis of language. Terrace argues that this theory explains Nim's inability to acquire words and, more broadly, the differences between human and animal communication. Why Chimpanzees Can't Learn Language and Only Humans Can is a masterful statement of the nature of language and what it means to be human.

Chapter One

NUMBERLESS GRADATIONS

ABOUT 3.8 billion years ago, the earth was just another lifeless speck in an infinite universe. Since that time, more than eight billion species of plants and animals have lived on our planet. In 1859, Charles Darwin proposed an ingeniously simple theory to account for this diversity.¹ He suggested that all existing species evolved by descent from a common ancestor by a process called natural selection.

Without any constraints, the size of the population of each species would increase exponentially. This did not happen because of steady competition for a finite set of resources in unpredictable environments. The prevailing rule was “Nature, red in tooth and claw.”² Because of predation or disease, some members of a species died before they could reproduce. The offspring of surviving members of any given species engaged in a new cycle of competition for survival. During that process, variations of genetic structure occasionally led to the formation of a new species. It is generally agreed that successful breeding is the criteria for distinguishing old and new species. Members of a new species cannot produce viable offspring by mating with members of an old species.

Thomas Huxley, one of Darwin’s strongest supporters, remarked, “How extremely stupid not to have thought of that!”³ Many other scientists agreed. During the following century, thousands of publications confirmed predictions of the theory of evolution. In 1973, Theodosius Dobzhansky, an eminent twentieth-century geneticist, concluded that “nothing in biology makes sense except in the light of evolution.”⁴

Language is the major exception. The theory of evolution has yet to explain it. Darwin tried, but his strongest statement about its origin was vague: He argued that language evolved from simpler forms of animal communication during the course of “numberless gradations” (italics added).⁵ We have yet to discover what those gradations are.

The challenge is to fill the seemingly unbridgeable gap between animal communication and language. As Noam Chomsky famously asked, how could language, a voluntary form of communication, have been selected from the involuntary grunts and screams of chimpanzees, our closest living ancestors?

Some of Darwin’s contemporaries also thought that the gap between animal communication and language was too large. For example, Max Müller, a professor of linguistics at Oxford, asserted that:

Language is the Rubicon which divides man from beast, and no animal will ever cross it… the science of language will yet enable us to withstand the extreme theories of the Darwinians, and to draw a hard and fast line between man and brute.⁶

Experiments on “ape language,” which I describe in chapter 2, support Müller’s conclusion that animals cannot acquire language. But he is wrong about evolutionary theory. There is simply no scientific alternative. What remains elusive are the numberless gradations that led from animal communication to language.

Some members of the European intelligentsia, Darwin included, naïvely argued that our ancestors’ first words were imitations of natural sounds: for example, the sounds of a dog barking, a snake hissing, a cat purring, a duck quacking, workers exerting themselves, and so on.⁷ In this view, language was just another example of animal communication, albeit more complex. The obvious and fatal problem with such “onomatopoetic” theories is that, even if true, they could account for only a miniscule portion of vocabulary. In response to the facile quality of the theories being offered, Société de Linguistique de Paris (in 1866) and the London Philological Society (in 1872) banned all further discussions of the origin of language shortly after the publication of the Origin of Species.

That did not deter a strong desire to alleviate what many thinkers regarded as the Achilles’ heel of the theory of evolution. Following Darwin, linguists, anthropologists, psychologists, neuroscientists, biologists, computer scientists, philosophers, and others have written about the evolution of language. This book continues that effort but with a novel approach. Instead of focusing on grammar, our most powerful linguistic ability, I ask a simpler but equally important question: Why are humans the only species that uses words, that knows that things have names, and that can use names conversationally?

Why opt for words instead of grammar? Quite simply, I changed my mind about their relative importance after the failure of Project Nim, the goal of which was to teach a chimpanzee to create a sentence. The failure of that project led me to realize that chimpanzees were not only unable to produce sentences, but they were also unable to produce words. I recognized that asking how language evolved was actually asking two questions: one about the origin of words, the other about the origin of grammar.

Understanding the origin of words is a much more difficult problem than it might first appear. Like walking, talking is so universal that it hardly seems worthy of our attention. Indeed, walking and talking are similar in that, with no obvious instruction, humans seem to develop both abilities on their own.

But they don’t. To see why not, imagine the proverbial story of a newborn infant who is raised alone on an island on which all of her basic needs are met. In theory, a robot could be programmed to satisfy those needs (for example, to provide food, drink, warmth, and so on). Would this infant learn to walk? Of course! But would she learn to talk? Other than the reflexive cries one would expect from an infant in distress, I can’t think of any reason that she would. And even if that infant exclaimed her needs, and had them satisfied, there’s no reason to think that she would be able to conduct the simplest type of conversation.

The moral of this thought experiment is that learning to walk is a product of maturation. Before an infant can learn to talk, however, she must experience certain affective and cognitive interactions with others, usually her mother. Those interactions are nonverbal and uniquely human. Affectively, an infant learns to share her emotions. Cognitively, she learns to share her perceptions of interesting events. Both types of sharing are conversational. When viewed from this perspective, the evolution of language depends on the evolution of the ability to converse nonverbally and verbally.

There’s no evidence that a nonhuman animal can engage in conversation. The signals that animals use to communicate encompass a variety of physical dimensions. Male primates often stare at other members of their group to assert their position in a dominance hierarchy.⁸ Some species release pheromones to attract mates.⁹ Certain varieties of fish generate electrical signals during courting.¹⁰

Birds sing to attract mates and to assert their territory. Superficially, it may seem that songbirds’ duets are conversational, but analyses of their duets have shown that the songs are innate and that the topic of their exchanges does not vary.¹¹ A songbird’s message doesn’t transmit any new information. As another example, foraging bees communicate the location and quality of nearby food to their hive mates by engaging in elaborate “round” and “waggle” dances. Those dances, however, are not learned and are immutable. A bee cannot substitute a new dance to communicate the source and the quality of nearby nectar.¹² The same is true of the distinctive alarm calls that vervet monkeys produce when they perceive particular predators.¹³ However fascinating, such innate signals are few in number and rarely exceed thirty different types in a particular species.¹⁴ By contrast, human mothers and infants engage in dialogues that reflect how they each perceive the other’s emotional and mental state.

Foolishly, I never considered the role of conversation when I began Project Nim. It was generally assumed that all communication was conversational, including the communication described in the initial reports of ape language experiments. That assumption, which was clearly wrong, reflected my then anticognitive bias. Having been trained as a behaviorist, I was taught not to ask questions about the emotions and mental states of a speaker and a listener.

Psychologists are still struggling to answer these questions; but, as we shall see in chapter 4, progress has been made by supplementing behavioral methods rather than by replacing them. To understand those approaches, we must first understand the origins and limitations of behaviorism, the dominant theory of psychology during the first half of the twentieth century. For that, we must digress to the beginning of comparative psychology.

COMPARATIVE PSYCHOLOGY

In 1865, Herbert Spencer, a disciple of Darwin, broadened the theory of evolution by arguing that natural selection applies to behavior with the same logical force that it does to anatomical structures.¹⁵ Just as a creature’s sensory organs, brain size, or skeleton can be said to be adaptive, so can particular types of behavior. For example, a species can be distinguished by the degree to which its behavior is instinctive (unlearned) or the degree to which a particular behavior is susceptible to different types of classical or operant conditioning.

Spencer’s insight gave rise to comparative psychology, an area of psychology that examines similarities and differences in the behavior of different species. For obvious reasons, comparative psychologists were also called behaviorists. Of particular interest were similarities between human and primate intelligence that pertain to the evolution of language.

Research on insight in apes provides an instructive example. In a famous experiment by Wolfgang Köhler, Sultan, a captive chimpanzee, was faced with the problem of obtaining a banana that hung from the ceiling of his cage. To obtain the banana, Sultan stacked some wooden boxes that he played with, and then stood on top of the stack to reach the banana. Because he was not explicitly trained to do so, Köhler described Sultan’s “insight” as a remarkable intellectual achievement, one that deviated sharply from trial-and-error learning.¹⁶

Most comparative psychologists reject insight and other mentalistic concepts because evidence of those concepts cannot be observed directly. Instead, they base their theories exclusively on objectively defined stimuli and responses, a credo of behaviorism.

Many years after Sultan became a textbook example of an intelligent ape, a replication of Köhler’s experiment showed that Sultan’s performance actually required trial-and-error learning. Two groups of chimpanzees were trained under conditions identical to th...

Table of contents

• Cover 
• Series Page
• Title Page
• Copyright
• Dedication
• Contents 
• Preface
• Prologue
• 1. Numberless Gradations
• 2. Ape Language
• 3. Recent Human Ancestors and the Possible Origin of Words
• 4. Before an Infant Learns to Speak
• 5. The Origin of Language, Words in Particular
• Epilogue
• Notes
• References
• Index