Can a monkey, a pigeon or a fish reason like a person? It’s a question scientists have been testing in increasingly creative ways – and what we’ve found so far paints a more complicated picture than you’d think.

Imagine you’re filling out a March Madness bracket. You hear that Team A beat Team B, and Team B beat Team C – so you assume Team A is probably better than Team C. That’s a kind of logical reasoning known as transitive inference. It’s so automatic that you barely notice you’re doing it.

It turns out humans are not the only ones who can make these kinds of mental leaps. In labs around the world, researchers have tested many animals, from primates to birds to insects, on tasks designed to probe transitive inference, and most pass with flying colors.

As a scientist focused on animal learning and behavior, I work with pigeons to understand how they make sense of relationships, patterns and rules. In other words, I study the minds of animals that will never fill out a March Madness bracket – but might still be able to guess the winner.

Logic test without words

The basic idea is simple: If an animal learns that A is better than B, and B is better than C, can it figure out that A is better than C – even though it’s never seen A and C together?

In the lab, researchers test this by giving animals randomly paired images, one pair at a time, and rewarding them with food for picking the correct one. For example, animals learn that a photo of hands (A) is correct when paired with a classroom (B), a classroom (B) is correct when paired with bushes (C), bushes (C) are correct when paired with a highway (D), and a highway (D) is correct when paired with a sunset (E). We don’t know whether they “understand” what’s in the picture, and it is not particularly important for the experiment that they do.

One possible explanation is that the animals that learn all the tasks create a mental ranking of these images: A > B > C > D > E. We test this idea by giving them new pairs they’ve never seen before, such as classroom (B) vs. highway (D). If they consistently pick the higher-ranked item, they’ve inferred the underlying order.

What’s fascinating is how many species succeed at this task. Monkeys, rats, pigeons – even fish and wasps – have all demonstrated transitive inference in one form or another.

The twist: Not all tasks are easy

But not all types of reasoning come so easily. There’s another kind of rule called transitivity that is different from transitive inference, despite the similar name. Instead of asking which picture is better, transitivity is about equivalence.

In this task, animals are shown a set of three pictures and asked which one goes with the center image. For example, if white triangle (A1) is shown, choosing red square (B1) earns a reward, while choosing blue square (B2) does not. Later, when red square (B1) is shown, choosing white cross (C1) earns a reward while choosing white circle (C2) does not. Now comes the test: white triangle (A1) is shown with white cross (C1) and white circle (C2) as choices. If they pick white cross (C1), then they’ve demonstrated transitivity.

The change may seem small, but species that succeed in those first transitive inference tasks often stumble in this task. In fact, they tend to treat the white triangle and the white cross as completely separate things, despite their common relationship with the red square. In my recently published review of research using the two tasks, I concluded that more evidence is needed to determine whether these tests tap into the same cognitive ability.

Small differences, big consequences

Why does the difference between transitive inference and transitivity matter? At first glance, they may seem like two versions of the same ability – logical reasoning. But when animals succeed at one and struggle with the other, it raises an important question: Are these tasks measuring the same kind of thinking?

The apparent difference between the two tasks isn’t just a quirk of animal behavior. Psychology researchers apply these tasks to humans in order to draw conclusions about how people reason.

For example, say you’re trying to pick a new almond milk. You know that Brand A is creamier than Brand B, and your friend told you that Brand C is even waterier than Brand B. Based on that, because you like a thicker milk, you might assume Brand A is better than Brand C, an example of transitive inference.

But now imagine the store labels both Brand A and Brand C as “barista blends.” Even without tasting them, you might treat them as functionally equivalent, because they belong to the same category. That’s more like transitivity, where items are grouped based on shared relationships. In this case, “barista blend” signals the brands share similar quality.

Researchers often treat these types of reasoning as measuring the same ability. But if they rely on different mental processes, they might not be interchangeable. In other words, the way scientists ask their questions may shape the answer – and that has big implications for how they interpret success in animals and in people.

This difference could affect how researchers interpret decision-making not only in the lab, but also in everyday choices and in clinical settings. Tasks like these are sometimes used in research on autism, brain injury or age-related cognitive decline.

If two tasks look similar on the surface, then choosing the wrong one might lead to inaccurate conclusions about someone’s cognitive abilities. That’s why ongoing work in my lab is exploring whether the same distinction between these logical processes holds true for people.

Just like a March Madness bracket doesn’t always predict the winner, a reasoning task doesn’t always show how someone got to the right answer. That’s the puzzle researchers are still working on – figuring out whether different tasks really tap into the same kind of thinking or just look like they do. It’s what keeps scientists like me in the lab, asking questions, running experiments and trying to understand what it really means to reason – no matter who’s doing the thinking.

Olga Lazareva does not work for, consult, own shares in or receive funding from any company or organization that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.