In the 1970s, a young gorilla known as Koko gained worldwide attention for her ability to use human sign language. However, skeptics believe that Koko, and other animals that have “learned” to speak (including chimpanzees and dolphins), cannot truly understand what they are “saying” and that other species They claim they are trying to force them to use human language. Being physically present is futile.
“It’s anthropomorphic. We need to understand non-human communication in its own terms,” says a group of researchers keen to examine whether animals can engage in symbolic communication, at the University of British Columbia. says Karen Bakker, professor and fellow at the Harvard Radcliffe Institute for Advanced Study. Scientists are now using advanced sensors and artificial intelligence techniques to observe and decipher how a wide range of species, including plants, share information in unique ways of communicating. This field of “digital bioacoustics” is the subject of Bakker’s new book. The Sound of Life: How Digital Technology Brings Us Closer to the Flora and Fauna World.
Scientific American with Bakker on how technology helps humans communicate with creatures such as bats and bees, and how these conversations are forcing us to rethink our relationship with other species. talked.
[An edited transcript of the interview follows.]
Can you briefly tell us about the history of humans trying to communicate with animals?
In the mid-20th century, there were many attempts to teach human language to primate non-humans such as Koko. And those efforts were somewhat controversial. Looking back, one of the observations we have now (maybe less common back then) is that our approach was too human-centric. The hope then was to measure non-human intelligence by teaching non-humans to speak like we do. One of the terms used in this book is the concept of: environment, which is this notion of the living experience of the organism. if we are careful environment We don’t expect bees to speak human language, but we become very interested in their fascinating language, which is vibrational and positional. Sensitive to even the most nuanced. And that is the science of today. The field of digital bioacoustics has accelerated exponentially, revealing fascinating findings about communication across the tree of life, but now we are getting closer to these animals, saying, “They can talk like humans.” Can you do it?” But “Can complex information be communicated to each other? How do they do it? What’s important to them?” I think it’s a less human-centric approach.
To put it into perspective, I think it’s also important to recognize that there is an age-old tradition of “deep listening” to listen to nature. It’s an ancient art that is still practiced in an unmediated form. So when you combine digital listening with deep listening to open up a vast new world of non-human sounds and decipher them with artificial intelligence, I think we’re on the brink of two important discoveries. The first is non-human language. And that’s a very controversial statement, and we can dig into it.
What technologies are enabling these breakthroughs?
Digital bioacoustics relies on extremely small, portable, lightweight digital recorders, like the miniature microphones scientists have placed everywhere from the Arctic to the Amazon. You can attach these microphones to the backs of turtles and whales. You can put them deep in the ocean. [put them] At the highest peak, attach it to the bird. It can also record sound continuously 24/7, in remote areas that scientists cannot easily reach, even in darkness, without the disruption of introducing human observers into the ecosystem.
A large amount of data is generated by such equipment. That’s where artificial intelligence comes in. Because the same natural language processing algorithms that work so well in tools like Google Translate can also be used to detect patterns in non-human communication.
What are examples of these communication patterns?
In the bat chapter, where we talk about Yossi Yovel’s research, there are specific studies he oversaw. [nearly two] Egyptian flying foxes were reared and recorded for two and a half months … [their] Vocalization.His team then adapted and analyzed the speech recognition program [15,000 of] Algorithms associated specific sounds with specific social interactions captured in the video. Using this, researchers were able to classify the majority of bat sounds. In this way, other researchers such as Yobel and Jerry Carter were able to pin down that bats have a much more complex language than previously understood. To do. They actually distinguish between genders when communicating with each other. They have individual names, or “signature calls.” Mother bats speak to their babies in ways that correspond to their “mother tongue.” But whereas a human mother raises the pitch of her voice when speaking to her baby, the mother bat lowers it. This triggers the babble response in babies who learn to “speak” specific words and instructional signals as they grow. As such, bats are engaged in vocal learning.
This is a great example of how deep learning can derive these patterns. [this] Instruments, all these sensors and microphones, reveal something inaccessible to the naked human ear. Most of bats’ communication is ultrasonic, which is beyond our range of hearing, and since bats speak much faster than we do, we have to slow down and lower the frequency to hear them. You can’t hear like, but computers can. And, of course, the next insight is that our computers can talk to bats, too. [The software produces] Use specific patterns to communicate with bat colonies or beehives. This is what researchers are doing now.
How do researchers talk to bees?
Studying bees is interesting.a [researcher] The communication of a bee named Tim Landgraf, as mentioned earlier, is vibratory and positional. Body movements and sounds are important when bees “talk” to each other. Because computer vision can be used in combination with natural language processing, computers, especially deep learning algorithms, can follow suit. They have now perfected these algorithms to the point where they can actually track individual bees and determine how individual communications affect other bees. gives you the ability to decipher the It turns out that they have certain signals. [Researchers have given these signals] funny name. [Bees] Toot; they squeak. There are signals for ‘Quiet’ or ‘Stop’, and signals for loud ‘Danger’.they have plumbing [signals related to swarming] Begging and waving signals, they all direct collective and individual action.
Landgraf’s next step was to encode this information into a robot called RoboBee. Ultimately, after building seven or eight prototypes, Landgraf came up with a “bee” that could enter the hive. Landgraf’s bee robot can therefore tell other bees to stop. It can also do more complex things, the very famous waggle dance. Because we put the nectar source in a place that the bees in the hive have never visited, tell the robot to tell the bees where the nectar source is, and then see where the nectar source is. Whether the bees fly well there. And indeed they do. That’s an amazing result.
This raises many philosophical and ethical questions. One can imagine such a system being used to protect bees, for example instructing bees to fly to safe nectar sources rather than polluted ones containing high levels of pesticides. It is conceivable that this could be a tool for domesticating previously poorly domesticated wild species, or for attempting to control the behavior of other wild species. And insights into the degree of sophistication and complexity of non-human language raise some very important philosophical questions about the uniqueness of language as a human capacity.
How is this technology impacting our understanding of the natural world?
The invention of digital bioacoustics is like the invention of the microscope.when [Dutch scientist Antonie] van Leeuwenhoek began peering into the microscope and discovering the microbial world…and it laid the foundation for countless future breakthroughs. Microscopes have enabled humans to see new things with both their eyes and their imagination. The analogy here is that digital bioacoustics combined with artificial intelligence is like a global hearing aid that enables new hearing with both artificially enhanced ears and imagination.This slowly opens our minds not only to the wonderful sounds that non-humans make, but also to a fundamental set of questions about the so-called separation between humans and non-humans, their relationship to other species. and [it’s] It also opens up new ways of thinking about conservation and our relationship to the planet. It’s pretty deep.
