Half a century ago, one of the hottest questions in the scientific community was whether humans could teach animals to speak. Scientists have tried communicating with apes using sign language and trained parrots to use the growing vocabulary of English.
The work quickly attracted media attention and controversy. The study lacked rigor, and it's possible that what appeared to be animal communication was just wishful thinking, with the researchers unconsciously telling the animals to respond in a certain way. critics argued..
In the late 1970s and early 1980s, this research fell out of favor. “The whole field completely collapsed.” says Eileen Pepperberg, a comparative cognition researcher at Boston University. He became known for his work on a gray parrot named Alex.
Today, advances in technology and a growing awareness of the sophistication of animal minds have sparked renewed interest in finding ways to bridge species gaps. Pet owners are teaching their dogs to press the “talk button” and zoos are training apes to use touch screens.
In a careful new paper, a team of scientists outlines a framework for evaluating whether such tools have the potential to give animals new ways to express themselves. Jennifer Cunha, a visiting scholar at Indiana University, said the study “aims to move past some of the things that have been controversial in the past.”
The paper, to be presented at a scientific conference on Tuesday, focuses on Cunha's parrot, an 11-year-old Goffin cockatoo named Ellie. Since 2019, Cunha has been teaching Ellie how to use an interactive “speech board.” The app includes more than 200 illustrated icons that correspond to words and phrases such as “sunflower seeds,” “happiness,” and “feeling good.” hot. “When Ellie presses an icon with her tongue, a computerized voice reads the word or phrase.
In the new study, Cunha and her colleagues did not determine whether Ellie's use of the speech board constituted communication. Instead, they used quantitative computational methods to analyze Ellie's icon presses, taking a closer look at whether the speechboard had what they called “expressive and rich potential.”
“How can we analyze expressions to see if there is intent and space for communication?” Cunha said. “And secondly, do her choices give us an idea about her values, what she finds meaningful?”
Scientists analyzed about 40 hours of video footage collected over seven months of Ellie using the speechboard. We then compared her icon presses to several simulations of her virtual speechboard users selecting icons randomly.
“They all ended up being very different from the real data in multiple ways,” said Nikhil Singh, the MIT doctoral student who created the model. “This virtual user we had couldn't quite capture what the real Ellie did when she used this tablet.”
In other words, it seems that whatever Ellie is doing, she's not just randomly hitting icons. The design of the speechboard, including the brightness and position of the icons, also did not fully explain Ellie's choices, the researchers found.
Determining whether Ellie's selection was random “is a very good place to start,” he said. Federico Rossano, a comparative cognition researcher at the University of California, San Diego, was not involved in the study. “The problem is that the randomness is very low.”
Dr. Rossano said just because Ellie wasn't randomly hitting the icons, it didn't mean she was actively and intentionally trying to communicate her true desires and feelings. She may have simply repeated the sequences she learned during training. “It’s like a vending machine,” he said. “You can learn how to press a series of numbers to get a specific reward. It doesn't mean you're thinking about what you're doing.”
To investigate the possibility further, the research team looked for signs of what they called “corroboration.” If Ellie selected the apple icon, did she eat the given apple? If she selected the reading-related icon, did she read the book for at least one minute of her time?
“If you give something to a bird, it's likely to throw it or touch it,” Cunha said. “But for us, it was whether she had a hand in it.”
Not all of Ellie's choices can be evaluated in this way. For example, it was impossible for researchers to determine whether she was truly feeling happy or hot at that moment. However, of the approximately 500 icon presses that could be evaluated, 92 percent were corroborated by Ellie's subsequent actions.
“It's clear there's a good correlation there,” said Dr. Pepperberg, who was not involved in the study.
However, she said additional tests are needed to prove that Ellie really understands the meaning of the icons, and suggested researchers intentionally try bringing Ellie the wrong object. To see how she would react. “This is just another control to make sure the animal really understands what the label represents,” Dr. Pepperberg said.
Finally, the researchers sought to assess whether the speechboard served as a means of enriching Ellie by analyzing the types of icons she most frequently selected.
“If it's a means to an end, then what is the end?” said Rebecca Kleinberger, author of the paper and a researcher at Northeastern University who studies how animals interact with technology. he said. “It seems like there was a bias against social activities and activities that meant continuing to interact with management.”
The researchers found that about 14 percent of the time, Ellie chose the food, drink, and snack icons. However, about 73 percent of the activities she selected were socially or cognitively enriching, such as playing games, visiting other birds, or simply communicating with Cunha. . Ellie also began using her board for speech 85 percent of the time.
Amalia Bastos, a comparative cognition researcher at Johns Hopkins University and not an author on the paper, said, “Ellie the parrot consistently interacted with the device, and the device remained attractive to her over several months, reinforcing it.'' This suggests that he continued to do so.”
Research has limitations. Outside experts said there are limits to what scientists can extrapolate from a single animal, and it's difficult to rule out the possibility that Cunha was unconsciously telling Ellie to respond in a certain way. . But scientists also praised the researchers' systematic approach and modest claims.
“They're not saying, 'Can parrots talk?'” Dr. Rossano says. “They're saying, 'Can we use this for enrichment?'”
Dr. Bastos agreed. “This study is an important first step,” she said. This is also an example of how the field has changed for the better since the 1970s.
“Researchers currently working in this area are not putting the same assumptions on the table,” Bastos said. “We don't expect animals to understand or use language in the same way as humans.” Instead, scientists are focusing on “the welfare of captive animals and their relationships with their caretakers.” She added that she is interested in using communication tools to “improve
