A study published June 26 revealed that Native Amazonians, Americans and monkeys show similar thinking patterns in the ability to represent sequential patterns.
In experiments with 100 participants from different age groups, cultures and species, researchers tested what makes human minds capable of learning language. The study found that the human capacity to generalize structures may be present in monkeys as well.
According to Steven Piantadosi, a researcher for the experiment and UC Berkeley assistant professor of psychology, the scientists were dissatisfied with prior experiments, which did not have definite results.
“The idea of this study was to give (participants) data that was ambiguous about whether they should use recursion or not, and study how they generalize, which we think is hard to explain with anything other than recursive generating processes,” Piantadosi said in an email.
In the study, scientists tested three macaque monkeys and three groups of humans: 50 preschoolers and kindergartners, 10 adults from the United States and 37 individuals from a group of Native Amazonians called the Tsimané, who live as forager-farmers in Bolivia.
“We needed to design a procedure that could be implemented with only small changes across widely different participants,” said Stephen Ferrigno, a postdoctoral research fellow at Harvard University, in an email.
In each trial, participants were shown images of four brackets randomly placed on a computer screen or, in the case of the Tsimané, on paper cards and were trained to touch them in a way that would produce a center-embedded sequence, such as “( [ ] ).”
The monkeys were given juice if they got the sequence right, and they were then tested on what they had learned and if they could generate the same sequences without feedback.
The first basic finding was that, as expected, all humans quickly learned to generate center-embedded sequences, with a possible increasing tendency with more education or more reading and writing experience.
After extended training and more examples, two of the three monkeys learned to generate sequences as well and could generalize patterns to new lists.
“The capacity to represent nested sequences is present in an animal that can never and will never learn language,” Ferrigno said in the email. “This suggests that this ability is more evolutionarily ancient than language and could have been a precursor to the development of human grammar.”
The monkeys in the study, however, did make mistakes. According to Sam Cheyette, a campus graduate student in Piantadosi’s lab, the researchers found evidence that working memory or attentional limitations may play a role in the monkeys’ subpar performance and hinder them from learning complex forms of communication.
The next step in this research program, according to Ferrigno, is to investigate how these recursive sequences are represented in the mind.
“We are also working on follow up experiments testing how people with Specific Language Impairment represent complex sequences,” Ferrigno said in the email.
He added that the long-term goal is to gain a deeper understanding of grammar-based language impairments to create targeted interventions.