New research has shown for the first time that animals are capable of the deductive aspects of a form of reasoning by exclusion through a decision-making task completed by primates, indicating that verbal language is not a prerequisite for this type of logic.

In a paper that was published Jan. 25 in Psychological Science, a team of researchers from Harvard University studied the capacity of olive baboons to use a nonverbal disjunctive syllogism to find hidden grapes. They used a four-cup task to test whether the primates could rule out alternative food locations, and found that they were successful at finding the grapes nearly 75% of the time.

A disjunctive syllogism is a form of logical inference that is reasoning by exclusion. An example of one is as follows: If it is true that someone will eat soup or they will eat salad, and it is also true that they will not eat soup, then it can be inferred that they will eat salad. 

“Logic is a critical aspect of early childhood learning, and studying how monkeys perform on these tasks is one way of looking at what’s possible in non-humans, and also what’s possible in a mind that can’t talk,” Stephen Ferrigno, a postdoctoral fellow at Harvard University and co-author of the paper, told The Academic Times.

There has been debate over whether logic words such as “or” and “and” are necessary for reasoning through these types of logical problems, Ferrigno said. Other logical reasoning studies done with children have shown that they cannot reason through alternative possibilities before 3-4 years old, when their language is more developed.

Therefore, if this form of reasoning is dependent on verbal language, then it shouldn’t be possible in non-human animals. Primates were used in this study because they have short-term visuospatial memory abilities similar to or better than those of 4- to 5-year-old children. 

“How monkeys perform on a task really gets at the question of whether these logic words are needed, because monkeys don’t have these logic words and they’re never going to learn them,” Ferrigno said.

“So if we can find evidence that a monkey can reason through a logic problem, that suggests that learning these logic words are not needed to be able to represent the underlying logical problem,” he continued.

Previous research has shown that non-human animals are capable of reasoning by exclusion, which is one aspect of the disjunctive syllogism. But according to the current paper, until now it was unknown whether non-human animals are capable of the deductive aspects of a disjunctive syllogism.

Ferrigno and his team conducted their experiment with four adult olive baboons, giving them disjunctive syllogism problems that they could solve to earn grapes. The baboons were shown four cups split into two sets. Two cups were covered, and a grape was hidden in one of those two. Then the other two cups were covered, and a grape was also hidden in that set.

The primates had two choices to find a grape. If they chose an empty cup on their first guess, and if they were able to reason using disjunctive syllogism, then they should have been able to infer that the grape was in the other cup within the same set. This is how the baboons consistently responded — in the trials when this occurred, the four primates made this conclusion 59% of the time.

When the location of the grape could still be in the first set, the baboons also had higher confidence in their second choice. They were more likely to choose quickly, even before the researcher indicated it was time for them to make a choice.

If the primates had happened to find a grape on their first choice, then the logical response for their second choice would be choosing a cup from the other set. This again was the majority finding, and the primates correctly made the switch to the other set 66% of the time.

The baboon participants in the study were named Olive Oil,  Pepperella, Jefferson and Sabina. Three of the four chose the cup that had to contain the grape more often than the determined chance amount of 33%. Olive Oil made the correct choice 92% of the time, Pepperella was 75% correct and Jefferson was 45% correct. Only Sabina fell below the chance amount, with 21% correct choices.

Studies of this nature have typically been done with a two-cup design, which cannot test for the deductive reasoning aspect, the authors said. The second set of cups in the current study have a 50% chance of having a grape, which allows the researchers to compare whether the primates know that the remaining cup must have the grape, or whether they're simply choosing other locations by default. 

Overall, the primates were capable of reasoning through a logical problem. When it was possible to infer the location of the remaining grapes, they were able to do that and choose the correct cup, the researchers noted.

The study was modeled after one that Ferrigno’s current advisor at Harvard ran with children between 2-4 years old. Children two-and-a-half years old could not reliably infer which location must contain a hidden object, and performance in that study did not reach above 50% until the age of 4 years.

“This failure of young children could be due to a lack of nonlinguistic modal concepts,” the authors said in the paper. “But the success of monkeys on this task suggests that monkeys not only are capable of representing a possible outcome ... but also can compare multiple possibilities.”

Ferrigno noted that this is one case study in the broader topic of logical reasoning, and there should be further research done to understand what aspects of logical reasoning are possible in non-human animals and when these abilities develop in children.

The study, “Reasoning Through the Disjunctive Syllogism in Monkeys” was published Jan. 25 in the Psychological Science journal. Stephen Ferrigno, of Harvard University, Yiyun Huang, of Yale University, and Jessica Cantlon, of Carnegie Mellon University, served as co-authors of the paper.