Studied chimps were capable of differentiating between genuinely new information and redundant details (Patrick Rolands/Shutterstock)

In A Nutshell

• In different experiments testing how they use information to make decisions, chimpanzees didn’t just react to sights and sounds, they compared how strong different clues were before deciding what to believe.
• They showed “rational belief revision,” meaning they changed their minds when better evidence appeared and stuck with their choice when earlier evidence was stronger.
• The apes could tell the difference between new and repeated information, and they recognized when earlier clues were undermined by later ones.
• Their decisions matched mathematical models of rational reasoning, suggesting that the roots of logical, evidence-based thinking run deep in our shared evolutionary history.

Chimpanzees can weigh and rationalize conflicting information before deciding what to believe. That’s the main conclusion of research that tested whether our closest living relatives can do something scientists call “rational belief revision,” a thinking skill previously thought to set humans apart.

Researchers at UC Berkeley, the University of Portsmouth, and other institutions designed a series of clever experiments with 15 to 23 chimpanzees at Ngamba Island Chimpanzee Sanctuary in Uganda. The chimps watched as food was hidden in boxes, then received different types of clues about where to find it. Sometimes they saw the food directly through a glass window. Other times they only heard it rattling around inside a box or saw traces of it left behind.

After the chimps chose a box based on one piece of evidence, researchers gave them a second, contradictory clue pointing to a different location. The question was whether the apes would stick with their first choice or change their minds, and more importantly, whether they’d make that decision in a smart, strategic way.

The chimps didn’t flip-flop randomly as one might expect. They stayed committed to their initial belief when the first evidence was strong and the contradicting evidence was weak. But when the situation reversed and weaker evidence came first, they readily changed their minds after receiving stronger proof pointing elsewhere.

Chimpanzees Use Rational Thinking, Not Random Guessing

In one experiment, chimps heard food shaking in one box, then saw food directly in another box through a clear panel. Most switched their choice to the box where they’d seen the food, recognizing that visual proof beats auditory hints. In another version, they saw food traces near one box, then heard rattling in another. Again, they favored the stronger evidence, in this case the sound of actual food over mere leftovers.

Study authors wanted to know if chimps were truly thinking about evidence or just responding to whatever seemed flashiest. So they ran a test with three boxes: one with strong evidence, one with weak evidence, and one with no evidence at all. After showing the chimps all three options, they removed the box with strong evidence. If the animals were only paying attention to the most obvious cue, they should have chosen randomly between the two remaining boxes. Instead, they consistently picked the box supported by weak evidence over the one with none, suggesting they were tracking more than one possibility all along.

The researchers point out that changing your mind based on evidence strength requires keeping track of not just what you know, but how you know it.

Telling Real Information From Noise

Perhaps most impressively, the chimps could tell the difference between genuinely new information and redundant evidence. When researchers shook a box twice, the chimps recognized this as the same piece of evidence presented again. It didn’t make them more likely to pick that box. But when researchers dropped two separate pieces of food into a box one after another, creating two distinct sounds, the chimps treated this as accumulating evidence and were more likely to revise their choice.

The animals changed behavior when their original evidence was undercut. In one test, after a chimp chose a box where it had seen food through a window, researchers revealed that the window actually showed a picture of food, not real food. This undercut the earliervisual cue; chimps treated it as weaker and were more willing to switch boxes. When researchers revealed an empty window instead of a picture, which didn’t change anything about what they’d seen, the chimps mostly stuck with their choice.

In both experiments, chimps got two different clues about where food was hidden, then had to choose twice. The key question: Would they change their minds based on which clue was stronger?

Experiment 1: In some trials, chimps first heard a box being shaken (weaker clue), made a choice, then saw food directly through a clear window in the other box (stronger clue), and chose again. In other trials, they got the clues in reverse order—seeing first, then hearing.

Experiment 2: In some trials, chimps first saw leftover traces of food near a box (weaker clue), made a choice, then heard food rattling inside the other box (stronger clue), and chose again. In other trials, the order was flipped.

The arrows show when researchers rotated or moved boxes to reveal the visual clues. To keep the chimps from gaming the system, researchers also mixed in some trials where they only got one clue and one choice—so the animals never knew whether their first or second pick would be the one that counted. (Credit: Schleihauf et al / Science)

Why This Matters for Animal Intelligence

Previous research showed that great apes could figure things out based on various types of clues. A hungry chimp might locate hidden food by listening to sounds, looking at visual hints, or noticing that food is absent from other spots. But those abilities could be explained by simpler mental shortcuts that don’t require actually reasoning about how trustworthy different sources of information are.

This study, published in Science, goes further. The chimps went beyond simply responding to evidence — they took time evaluating it. When deciding whether to stick with or abandon a belief, they compared the strength of competing pieces of information.

The research team used mathematical models to predict how a perfectly rational thinker should behave when faced with conflicting evidence of different strengths. The chimps’ behavior matched these predictions remarkably well, not just overall but for nearly every individual animal tested.

These results show that the ability to weigh evidence and change your mind accordingly likely has deep evolutionary roots among great apes.

The study also addresses a longstanding debate about whether animals can truly think about their own thinking, or if they just happen to behave in ways that look smart. Critics have argued that when chimps search for more information or skip difficult tasks, they’re not really reflecting on their own knowledge—they’re just reacting to uncertainty in ways that mimic deeper thought.

But the experiments in this study make that skeptical explanation harder to accept. Recognizing when a piece of evidence has been proven wrong by later information requires understanding the relationshipbetween evidence and belief, not just reacting to whatever’s in front of you.

In the wild, chimpanzees face complex social problems and ecological challenges that benefit from smart decision-making. A chimp trying to figure out where ripe fruit might be located, or whether a particular sound signals danger, needs to pull together multiple sources of information and decide which ones to trust. Being good at this probably helps them survive.

---

Paper Summary

Methodology

Researchers conducted five experiments with chimpanzees at Ngamba Island Chimpanzee Sanctuary in Uganda. Between 15 and 23 chimps participated depending on the experiment: 15 in the first two experiments, 23 in Experiments 3 and 4, and 22 in the final one. In the main experiments, researchers hid food in one of two boxes, then gave the chimps two different types of clues pointing to different locations. The types of evidence included seeing food directly through glass, seeing traces the food left behind, or hearing food rattle in a box. Chimps made an initial choice after the first clue, then a second choice after getting contradicting evidence. The order of strong versus weak evidence was mixed up across different test conditions. One experiment tested whether chimps kept mental track of multiple options by using three boxes and removing the best option before letting them choose. Another experiment checked whether chimps could tell the difference between brand new evidence (a second piece of food dropping) and redundant evidence (shaking the same box twice). A final experiment looked at how chimps responded when researchers showed them their original evidence was misleading, like revealing that food seen through a window was actually just a picture. Researchers used statistical models to analyze the choices and compared results against mathematical models of rational thinking.

Results

In the main experiments, chimpanzees were much more likely to change their initial beliefs when weak evidence came first (followed by strong evidence) compared to when strong evidence came first (followed by weak evidence). After getting both types of evidence, chimps consistently chose the location supported by stronger evidence in all test conditions. Whether they saw the strong evidence first or second didn’t matter—they still ended up choosing it. Statistical modeling confirmed that chimps treated strong evidence as genuinely stronger than weak evidence, and this pattern held for almost every individual chimp. When the best option was removed in the three-box test, chimps picked the option with weak evidence much more often than the option with no evidence, rather than choosing randomly between them. Chimps changed their minds far more often when presented with new accumulating evidence compared to hearing the same evidence repeated. Chimps were much more likely to switch choices when shown that their previous evidence was misleading compared to when it was confirmed. Comparisons between different explanatory models showed that the rational thinking model best explained chimp behavior in most cases, outperforming explanations based on stubbornness, just picking the most recent option, or random chance. It performed equally well compared to a cue saliency model in the first two experiments.

Limitations

The study used a relatively small number of chimpanzees at a single sanctuary, which means the results might not apply to wild chimps or other captive groups. Between 15 and 23 individuals participated depending on the experiment, which is enough for the statistical tests but still represents a modest sample. The experiments required extensive training and repeated testing, and chimps did show some signs of learning over time in certain conditions (though this didn’t explain the overall pattern). The study only looked at food-finding tasks using visual and auditory clues. Whether chimps can revise their beliefs the same way in other situations, like social interactions or tool use, remains unknown. The research couldn’t completely rule out all possible simpler explanations, and one alternative model performed just as well as the rational thinking model in some comparisons. The study also couldn’t determine whether chimps consciously experienced awareness of their reasoning process or just behaved in ways consistent with rational thinking through other mental mechanisms.

Funding and Disclosures

Engelmann and Sanford received funding from the Leakey Foundation. Schleihauf received funding from the Leibniz ScienceCampus Primate Cognition. The authors reported no conflicts of interest.

Publication Information

Schleihauf, H., Sanford, E.M., Thompson, B.D., Zhang, S., Rukundo, J., Call, J., Herrmann, E., & Engelmann, J.M. (2025). Chimpanzees rationally revise their beliefs. Science, 30 October 2025. DOI: 10.1126/science.adq5229

About StudyFinds Analysis

Called “brilliant,” “fantastic,” and “spot on” by scientists and researchers, our acclaimed StudyFinds Analysis articles are created using an exclusive AI-based model with complete human oversight by the StudyFinds Editorial Team. For these articles, we use an unparalleled LLM process across multiple systems to analyze entire journal papers, extract data, and create accurate, accessible content. Our writing and editing team proofreads and polishes each and every article before publishing. With recent studies showing that artificial intelligence can interpret scientific research as well as (or even better) than field experts and specialists, StudyFinds was among the earliest to adopt and test this technology before approving its widespread use on our site. We stand by our practice and continuously update our processes to ensure the very highest level of accuracy. Read our AI Policy (link below) for more information.

Our Editorial Process

StudyFinds publishes digestible, agenda-free, transparent research summaries that are intended to inform the reader as well as stir civil, educated debate. We do not agree nor disagree with any of the studies we post, rather, we encourage our readers to debate the veracity of the findings themselves. All articles published on StudyFinds are vetted by our editors prior to publication and include links back to the source or corresponding journal article, if possible.

Our Editorial Team

Steve Fink

Editor-in-Chief

Sophia Naughton

Associate Editor