Do Crows Understand Cause and Effect?

In a Cambridge lab in 2009, a New Caledonian crow stood over a clear tube of water with a floating treat sitting just out of reach. The bird picked up a stone, dropped it in, watched the water rise, and dropped another. Researchers had expected roughly that much. What the crow did next is what rewrote the ceiling scientists had set on bird intelligence.

The test was built on an old fable about a thirsty crow that drops stones into a pitcher to raise the water level high enough to drink. Cambridge researchers made the fable real: a container of water, a treat floating just out of beak reach, and a pile of objects to choose from.

The short answer

Yes. Crows given the real version of Aesop's water-displacement puzzle consistently chose objects that would actually raise the water, picking heavy items over light ones and large over small, and ignoring a sand-filled tube that could never lift a floating reward. That pattern is genuine causal reasoning about volume, weight, and displacement, built step by step rather than produced by blind trial and error.

The controls that ruled out luck

The strength of the Cambridge result sits in what the researchers changed between trials. When crows chose between a water-filled tube that could lift the treat and an identical tube filled with sand that never could, they picked water. When heavy and light objects sat side by side, they picked heavy. When object sizes varied, they picked larger ones first. A bird guessing at random, or simply copying a memorized action, would not track all three variables correctly at once. A bird reasoning about cause and effect would.

Waiting for a better reward

A separate line of research tested something harder than tool logic: delay of gratification, the ability to pass up an immediate reward in favor of a better one later. Human children typically manage this around age four, and great apes can generally hold out for a few minutes. In a 2020 study, ravens, close corvid relatives, matched four-year-old children and beat great apes on the same task, waiting up to ten minutes for a preferred food item when the wait clearly paid off. Ten minutes is a long hold for a wild mind built around fast decisions.

Caching food while tracking who is watching

The most demanding territory is theory of mind, treating another individual as a holder of its own knowledge and intentions. The clearest corvid evidence comes from food caching. Crows and ravens hide food in hundreds or thousands of locations and retrieve it later, and some jays can hold roughly two hundred cache sites in mind at once. The sharper finding is what a bird does when it knows it was watched while hiding food: it often waits until the observer leaves, then quietly relocates the cache. When no one watched, the food stays put.

That relocation decision only makes sense if the bird is modeling what the watcher knows, this individual saw the spot, might come back for it, and should be given a reason not to bother. Tellingly, birds that have never stolen another's cache themselves take little evasive action when watched. It is the experienced thieves who move their food, projecting their own motives onto the observer, a form of reasoning built from lived experience rather than a fixed reflex.

Key terms

Causal reasoning

Working through cause and effect, such as how object weight or size changes a water level, rather than solving a problem by pure trial and error.

Theory of mind

Treating another individual as having its own separate knowledge, intentions, and beliefs, and adjusting behavior based on what that individual is likely to know.

Related field notes

For how this same reasoning shows up in tool making rather than tool choice, see more Crows Decoded field notes.

Crows Decoded covers the full run of caching and reasoning experiments, including what happens when a crow knows a fellow thief is watching.