Can animals learn human language

Not only children dream of talking with animals. Scientists experiment again and again with dolphins, dogs, parrots and, of course, monkeys - our closest relatives, with whom once upon a time we went different evolutionary ways. Linguist Sverker Johansson studied and described the most significant research and figured out whether we can understand each other with animals.

In Russian, his book “The Dawn of the Language. The way from monkey chatter to the human word "was published by the publishing house" Bombora ". Lifehacker publishes an excerpt from the first part.

---

The parrot has become popular due to its ability to learn human language. Or not, of course, the parrot only imitates sounds, without the slightest hint of understanding. The very word "parrot" (eng. Parrot) means exactly that.

The parrot has become adept at reproducing a variety of sounds, not only human speech, and sometimes with amazing accuracy. Of course, he did not develop his natural talent in order to learn how to speak like people. Parrots "parrot" for the same purposes that songbirds sing. Among nightingales, it is considered sexual to be able to produce a wide variety of trills, among parrots - to more accurately imitate a larger number of sounds.

Part of their social game is to imitate each other. The main thing is to surpass the opponent in the art of imitation. That's why parrots repeat everything they hear so often and with pleasure, especially in a social context. And people use it. If a parrot hears a human phrase many times in the process of communicating with a trainer, then he can then reproduce it quite accurately.

But do parrots learn language in the proper sense of the word? Hardly.

They usually memorize a few standard phrases, which they then repeat, obviously not understanding their meaning at all. And they never create new statements from memorized words.

The fact that they can reproduce human speech is amazing in itself. Few in the animal kingdom are capable of this. Among birds, except parrots, the habit of imitating what they hear is seen in hummingbirds and some songbirds, but most do not. In any case, no one has excelled in this art as much as parrots.

Among mammals, there are not many "imitators" at all, except perhaps some seals. Most animals cannot control their speech organs in such a way as to set them up to repeat the sounds they hear.

The abilities of monkeys in this regard are more than modest. For example, some individuals may repeat the sounds of others in order to adapt to the "dialect" of the pack in which they find themselves.

But people in the art of imitation are not much inferior to parrots and leave far behind all other mammals. We can imitate new sounds, and it gets better the longer and harder we practice. It works especially well with words. We easily repeat a new word that we have just heard. And children do learn to speak, constantly copying the speech of adults.

This ability is an indispensable condition for the existence of a spoken language.

If we did not know how to imitate someone else's speech, we would never learn to speak and would not be able to pass on the language from generation to generation.

At the same time, this talent is completely absent from our closest relatives, and therefore should have appeared somewhere in the process of evolution of the Homo sapiens species.

But why have we developed this ability at all? For the sake of language is the first answer that comes to mind. And then there's the chicken-and-egg problem.

The fact is that there is no distant future for evolution: certain qualities do not develop just because they will be useful in the future. And if the ability to imitate is necessary for the appearance of a language, then at the time of its occurrence it should already have been. But in this case, there were other reasons for its appearance.

For some birds, imitating the sounds of the world around them is one way to enrich their singing repertoire. Parrots do this for no apparent practical purpose. Perhaps in this way they expect to make acquaintances or gain influence. It is ultimately about new possibilities of copulation. What if the human talent for imitation has a similar origin? Perhaps, in our distant ancestors, the ability to imitate other animals influenced the social status? We have no evidence for this hypothesis.

Scientists drew attention to the imitative abilities of modern man, not related to linguistic goals. Hunters, mushroom pickers and other lovers of the forest often imitate the sounds of animals both on the hunt and later, talking about it. In conditions where there was no language, this ability could be of great importance, say, when planning a joint hunt. And this is one of the possible reasons for the development of "imitative" talent in a person.

• • •

- Bring the striped ball!

The white collie rushes to the end of the garden, where there are several balls and other toys, and returns with a striped ball.

Well done, smart dog. Now bring the duck.

For a while, the collie sorts through the toys, puzzled, but finally stops at the yellow plastic duck.

- Fine! Biscuit?

- Wow!

The dog grabs a treat, lies down next to the owner and, happy, chews.

• • •

How far do the linguistic abilities of our smaller brothers extend? Many of us have tried to teach human language to animals, with more or less varying success.

One thing is clear to anyone who has tried teaching horses, dogs, and other pets—they can be trained to understand certain verbal commands. Dogs learn the "sit" command without any problems. and after some training, they learn by ear to distinguish this word from others. In extreme cases, we can reinforce the order with a gesture. Sit on a chair when we say "sit", or get up from a chair, giving the appropriate command.

Many mammals are able to learn this, even if it works better with some animals than with others. It is more difficult to train a cat to sit on command than a dog. And it's not about intelligence, as my experience with cats tells me. Just following orders is really not a cat thing.

But the fact that a dog can adequately interpret our words does this mean that it understands human language? Well... at least it's a very limited understanding. The dog distinguishes between the words of different commands as long as it knows what it should do, say, at the word “sit”. If the words are related to food and feeding, there are no more problems with interpretation.

Among the dogs, there are especially talented ones who are able to learn hundreds of words, choose the right one from a pile of toys and bring it to the owner. But even in this case there can be no question of a full understanding of the language.

Animals just remember some words and associate each of them with a certain action.

There is nothing to indicate that the dog has any understanding of grammar. She just recognizes a certain keyword, no matter what the owners think about their pet, and reacts to it with a very specific action. Or reacts to our behavior with a certain action, for example, when we sit down, ordering her to sit, or filling the bowl with food. Nothing - alas - points to more.

The linguistic abilities of a person allow him to reason about what is not here and now, and in this direction, none of the dogs has so far seen any progress.

• • •

Two creatures are sitting at a table, on which a bunch of different little things are piled up, mostly children's blocks and balls of different colors.

“Give me the red die,” says being #1.

Creature #2 draws a red die from the pile and hands it to creature #1.

How many green balls are there? asks the first being.

“Three,” replies the second. — I want a nut.

Creature #2 gets a nut. No. 1 continues:

How many blue toys are there?

— Two.

No. 2 placed a blue ball and a cube of the same color in front of No. 1.

What are those green toys? asks #1.

“These are green balls,” answers No. 2.

- What a fine fellow you are! Here's another walnut for you.

• • •

Of course, dogs cannot speak humanly. Anatomically, their vocal apparatus is not adapted to the sounds of human speech, and dogs cannot control the vocal organs so that they can emit anything other than barking, growling or whining. The hero of the above dialogue is a parrot answering a man's questions. This bird, as we have already noticed, perfectly reproduces human speech.

But this parrot does not just imitate, he seems to use the language "for real", that is, he understands the questions and gives reasonable answers to them. The bird's name is Alex, and she has been trained by Irene PepperbergIrene Pepperberg, in addition to numerous articles, wrote the book "Alex and Me" about her pet. This is a non-fiction biography of Alex. Her other work, Teaching Alex, is a more formal overview of what the talented parrot could do. Alex died in 2007 at the age of 40. It is probably the only bird whose obituaries were published in newspapers such as The Economist and The New York Times. The dialogue above is my compilation of real dialogue replicas given in books by Irene Pepperberg. I took the liberty of showing off Alex's talents. The real dialogues with Alex are much longer and contain a lot of things that we would be wiser to skip.. Alex not only knows a lot of words, he uses them as if he understands the meaning. Can answer many questions about the shape, color and number of objects. If you ask him: “How many green balls are there?”, he will answer: “Three”, while on the table, in addition to three green balls, there are three more red and one more green cubes. And if you ask Alex: “What is that green there?” - pointing to the green ball, he will answer: “Ball”.

It is difficult to explain this in any other way than the fact that Alex understands human speech. In any case, he knows many concepts denoting different objects, color, shape and quantity. And his linguistic abilities are enough to put these concepts into words.

At the same time, Alex did not master the language enough to be able to maintain a general conversation on other topics than those that he was specially taught.

Nevertheless, Alex's achievements are impressive. Especially considering that we are talking about a creature whose brain is the size of a walnut. Despite this, he managed to master some part of the human language, and it remains to be seen to what extent Alex understood grammar.

The results of numerous attempts to teach other animals to speak are often much more modest. Parrots, perhaps, demonstrate the best abilities in this direction and can pronounce words almost like people.

Almost all experiments of this kind with monkeys can be considered unsuccessful. Monkeys cannot control their "speech" organs enough to reproduce human sounds and put them into words.

This also applies to chimpanzees raised in human families as adopted children, along with human “brothers” and “sisters”. A classic experiment was conducted in the 1930s in the United States, and a young chimpanzee was at first in no way inferior to a human child, except... language. Gua, that was the name of this chimpanzee, understood most of what was said to her, but at the same time could not extract a single more or less understandable word from her throat.

Instead, she responded with the usual monkey sounds, which, however, she adapted to connect in their own way and use in new contexts, but all this did not even remotely resemble a human speech.

On the other hand, language does not necessarily consist of sounding words, but nevertheless remains a language. And since it was precisely the reproduction of sounding speech that turned out to be an insurmountable barrier for monkeys, the attempts of researchers spread to non-verbal languages. A series of experiments since 1960 have used sign language or various artificial languages, when, for example, pressing a key or pointing to a symbol on the board meant pronounce the word. And classes with monkeys with the help of these improvised means were indeed much more successful.

Animals learned to use some "words" without problems and in the correct context.

The chimpanzee Washoe (1965–2007) excelled in her experimentation with sign language. The idea was the same as with Gua. Washoe grew up in a human environment, riddled with language. The only difference is that it was sign language. Washoe learned several hundred signs of Amslan, a language for the deaf spoken in the United States, and used them correctly in the right situations. In addition, she could combine a number of gestures into a perfectly reasonable statement.

Another experiment with sign language drew a line under many works on this issue. His hero was the chimpanzee Nim Chimpsky. Nim learned sign language in the same way as Washoe, but rather in a laboratory setting, where many scientific tests were carried out that confirmed his achievements.

This experiment is considered rather unsuccessful. Nim managed to learn very few gestures, and he practically did not know how to combine them. Herbert Terras, who was responsible for this work, concluded that chimpanzees have no aptitude for language, let alone grammar. The scientist reproached his predecessors for not being objective enough and interpreting the results of the experiments too optimistically.

In particular, Terrace pointed out, the effect of Clever Hans was not sufficiently taken into account.

• • •

Clever Hans is a horse who lived in Germany a hundred years earlier and became famous for his mathematical abilities. The owner of Clever Hans made good money on his talents. The horse could be asked any arithmetic problem, and he tapped out the answer with his hoof. For example, when asked about the square root of 25, there were five taps.

In the end, a psychologist was found who was suspicious of the equine genius and spent time with the animal an experiment that showed that Clever Hans can’t count at all, but he perfectly reads human emotions.

If you ask a question and the horse starts thumping, then you involuntarily tense up when he approaches the correct number. Clever Hans was only observant: by the expression of the face or posture of the questioner, he caught signs of tension or relaxation and stopped knocking at the right moment. When Clever Hans didn't see anyone who knew the right answer, he couldn't solve the simplest problem and continued to beat with his hoof until he was stopped.

This is the effect of Clever Hans.

Animals that are taught something often demonstrate something completely different from what people think, but they capture the most insignificant signs in the behavior of trainers and experimenters, on the basis of which they do what they they are waiting.

This factor must also be taken into account when teaching sign language to monkeys, since the trainer communicates closely with the animal and can give him a lot of unwitting clues on how to get a reward.

To protect against the Clever Hans effect, it is important that the animals in the experiment do not have visual contact with those who may unconsciously suggest the correct answer.

Until a certain point, this factor was practically not taken into account in experiments with chimpanzees, so it cannot be ruled out that, for example, Washoe acted on the same principle as Clever Hans. Only with Nim Chimpsky, the researchers became more careful, and the results immediately deteriorated. Many researchers have come to the conclusion that linguistic studies with monkeys are useless. Many, but not all.

In the 1970s, the experiments resumed, although after the fiasco with Nim Chimpsky, it became much more difficult to obtain funding. Gorilla Koko learned sign language and achieved even more impressive success than Washoe. According to her trainer, at the time of her death in 2018, Koko had mastered over a thousand gestures and applied them intricately in everyday life. But even in this case, there were reproaches that the effect of Clever Hans was not fully taken into account.

Dolphins also tried to learn languages ​​in many ways. And they showed good progress, both in the case of sounding human language, as well as sign language and specially developed based on whistling. In terms of understanding, they were not inferior to either monkeys or Alex the parrot. Rather, the difficulty is to get dolphins to express their thoughts in words that people can understand - with all the outstanding talent of these animals to imitate sounds.

Two chimpanzees, Sherman and Austin, participated in a different experiment, with different conditions and tasks. This experience deserves much more attention than it has received so far. Instead of placing the monkeys in a human environment, they were provided with a communication system suitable for "internal" monkey use, that is, for chimpanzees to communicate with chimpanzees.

Sherman and Austin sat each in their own room, each in front of their own keyboard with the same set of characters. They could not get to each other, but each saw on the screen which key the other was pressing. This allowed the monkeys to communicate using symbols with each other, which is much more interesting than answering stupid questions from bipeds.

Chimpanzees quickly adapted to using symbols to communicate messages to each other, and even learned to negotiate their new meanings.

When they were once given a new fruit for which there was no symbol on the keyboard, each held a treat in front of the screen, demonstrating to another, and then one of the chimpanzees selected a character on the keyboard and pressed key. So the monkeys agreed on how the new object would be designated in their language.

All this is very important, because this is how new words appear in the human language. A new concept arises, and a new word is required to designate it. Someone suggests or simply invents a word and starts using it. If others support it, the word sticks. This is the basis of the diversity and flexibility of human language, and within the framework of their "symbolic" language, Sherman and Austin did about the same thing.

Interestingly, in this situation, the chimpanzees used a linguistic ability that apparently never occurs in their natural habitat.

A turning point in the work with monkeys was the training of the bonobo Kanzi, born in 1980. Kanzi was small when his adoptive mother participated in an experiment in which she learned to communicate using symbols. Each symbol was located in a separate square on the computer screen or was attached with a magnet to an ordinary board, and Kanzi's mother had to carry on a conversation by pointing to the symbols.

Things weren't going very well. For a long time, my mother did not move anywhere. But one day, researchers (led by Sue Savage-Rumbaud) noticed that little Kanzi, who was in almost every lesson, learns much more than his mother. The attention of the experimenters shifted to the kid, who quickly learned the entire board with symbols.

Today he is not so small (each correct answer was rewarded with a candy: quite a few kilograms have been eaten over the years) and uses hundreds of characters in his “speech” without problems and understands spoken English at least as well as a two-year-old baby.

Kanzi quickly became popular among scientists and journalists alike. Now he is a key figure in a small group that includes monkeys and explorers. They conduct many joint experiments and communicate in everyday life using a board with symbols.

All experiments with Kanzi are carefully documented. The experimenters did their best to avoid the Clever Hans effect. Among other things, Kanzi was briefed over the telephone, in English, as usual. As soon as he hung up the phone, he began to perform the task. There was a man in the room with him (wearing earplugs so as not to hear the telephone conversation) who watched what Kanzi was doing and took notes. This man did not know what exactly Kanzi was entrusted with, and therefore could not tell him, as Clever Hans was told.

And the fact that Kanzi followed the instructions more or less correctly under such conditions indicates that he understood English. Of course, we are not talking about any language subtleties, but the instructions were not trivial. For example, Kanzi was asked to wash the carrots on the table in the kitchen and put them in a bowl in the living room. And the bonobo did the job flawlessly.

Kanzi could listen to the instructions on the phone and knew that there was a person on the other end of the line - it looks no less impressive.
Many stories survive of Kanzi's accomplishments in daily life, more or less documented. There is evidence that Kanzi was able to light a fire with matches and threw firewood into it, and then cooked an omelet on the fire.

Bonobo could make simple stone tools with a sharp edge and use them to cut the rope. Kanzi is said to have even played the Pac-Man computer game.

God bless him with Pac-Man, but the bonobos could do everything we thought Australopithecus could do, and much of what Homo erectus could do. On the other hand, no one has ever caught a chimpanzee in the jungle while he was frying an omelette or making a stone knife, not to mention Pac-Man. And again, we return to the fact that monkeys have hidden abilities that they do not use in the wild.

Kanzi's linguistic talents went far beyond the communications we can see in wild chimpanzees. But man also has many abilities that he does not use in the "state of nature", which in our case, apparently, means the life of a primitive hunter-gatherer.

Everything from solving differential equations to building a hydrogen bomb and writing this books - all these are human abilities that for the time being remained hidden and manifested themselves only in our days.

Alfred Russel Wallace, who came to the idea of ​​evolution and natural selection at the same time as Darwin, thought a lot about the problem of "higher mental abilities" of man. He came to the conclusion that natural selection does not explain how they arose, and that a qualitatively different, spiritual explanation is required here in addition to what is given within the natural sciences. This view is alive to this day among religious evolutionists. And in Wallace's day—and he published his ideas on the subject in the 1860s—it was supported by many scientists.

Within the framework of the natural-science picture of the world, such seemingly unnecessary abilities can be considered as a manifestation of a more general ability, which was used by our ancestors completely in other purposes.

Natural selection did not give rise to either mathematicians or engineers, but gave life to a biological species endowed with extraordinary cognitive flexibility, a highly developed ability to solve all conceivable problems that him life.

It was this ability that developed among primitive hunters and gatherers, since it allowed them to survive not only in the natural environment, but which they were originally adapted, but also in any natural conditions that are conceivable on our planet, from the arctic tundra to tropical atolls.

Those same abilities still help us deal with pressing problems, even if they are very different from those that our ancestors dealt with.

This, in particular, may explain why some of us can solve differential equations. The point is not at all that differential calculus so excited the minds of our ancestors. It's just that the intelligence that they managed to develop in themselves, we applied to differential calculus when it was required.

The same principles apply to the cognitive abilities of monkeys - much more modest than ours - including the ability to absorb certain aspects of human language.

It is of particular interest, including for the evolution of language, that some of the linguistic abilities of our closest relatives are hidden, that is, they do not appear in their natural habitat. Perhaps the same thing happened with our common ancestors 5-10 million years ago. Something was wrong with our ancestors that distinguished them from the ancestors of chimpanzees and contributed to the fact that language developed in us, but not in monkeys.

There must have been some essential difference between these two evolutionary lines, which, in in particular, it can serve as a good test for testing different theories of the origin of the language on credibility. A good theory should not only explain why we humans developed language, but also why it did not develop in chimpanzees or any other animals. This likelihood test is also called the "chimpanzee test".

The origin of language is one of the great mysteries of history. Scientists are still far from solving it, but with the help of archaeology, neuroscience, linguistics and biology, they can reject old hypotheses and put forward new ones. How did the language come about? Why do we say this and not otherwise? What was the first conversation about? Sverker Johansson tries to answer these and other questions in his book Dawn of Language.

To buy a book