Superstition-like behavior shows up across a surprisingly wide range of species, from pigeons and rats to monkeys and great apes. While no animal knocks on wood or avoids walking under ladders, many animals act on false associations between their own behavior and unrelated outcomes, and evolutionary models suggest this tendency is not a glitch but a built-in feature of adaptive learning. The story of animal superstition turns out to be less about quirky lab antics and more about how brains everywhere handle uncertainty.
Skinner’s Pigeons and a Famous Misunderstanding
The idea that animals can be superstitious traces back to B.F. Skinner’s 1948 experiment. He placed hungry pigeons in boxes that delivered food pellets at fixed time intervals, completely independent of anything the birds did. The pigeons developed repetitive behaviors: one bobbed its head, another turned in circles, a third swung its body like a pendulum. Skinner called this “superstitious behavior,” arguing that each bird had accidentally linked its random action to the food delivery, then kept repeating it because the food kept coming on schedule.
The interpretation became textbook psychology, but later research complicated it considerably. When researchers replicated the setup and looked more carefully, they found something different from what Skinner described. Rather than each pigeon developing its own unique quirky ritual, most pigeons under fixed-time food schedules converged on the same behaviors: pecking at the floor, orienting toward the food hopper, and other actions that look a lot like normal foraging. These are species-typical food-getting behaviors, the kind of thing a pigeon’s body is primed to do whenever food seems imminent, not random personal rituals reinforced by coincidence.1Behavioural Processes. Superstition revisited: Sex, species, and adventitious reinforcement The lack of idiosyncratic responding was a serious problem for Skinner’s original explanation. What looked like superstition in individual pigeons was more likely the bird’s innate feeding program kicking in because food was arriving regularly.
This does not mean animal superstition is a dead concept. It means the pigeon experiment, while iconic, was probably demonstrating something else: that animals have hard-wired behavioral packages that activate in the presence of food cues. True superstitious learning, where an animal forms a genuinely false causal belief and acts on it, needed to be demonstrated more carefully, and it has been.
Why Brains Are Built to See Patterns That Aren’t There
Before getting into which animals show superstitious behavior, it helps to understand why this tendency exists at all. An influential evolutionary model proposed that natural selection actually favors organisms that over-detect patterns, even at the cost of frequent false alarms. The logic is a version of “better safe than sorry.” An animal that hears rustling in the grass and assumes a predator is there will waste energy running from the wind sometimes, but it will survive the one time the predator is real. An animal that ignores the rustling saves energy on false alarms but gets eaten during the real one.2PubMed Central. The evolution of superstitious and superstition-like behaviour
The model showed that natural selection favors strategies that make frequent errors in causal assessment, as long as the occasional correct response carries a large enough survival benefit. In other words, a brain that produces lots of false positives (seeing patterns where none exist) beats a brain that produces false negatives (missing real patterns) in most ecological settings. The researchers concluded that superstition-like behaviors are “an inevitable feature of adaptive behaviour in all organisms, including ourselves.”2PubMed Central. The evolution of superstitious and superstition-like behaviour
A complementary model framed it as a trade-off between two kinds of mistakes: failing to exploit a real causal relationship versus wasting effort on a nonexistent one. Superstitions emerge as the inevitable byproduct of a learning system that tries to minimize both types of error simultaneously.3Animal Behaviour. The evolution of superstition through optimal use of incomplete information Another analysis put it even more bluntly: superstitious belief is an unavoidable consequence of any organism’s ability to learn from observed coincidences.4PubMed. Superstition and belief as inevitable by-products of an adaptive learning strategy The implication is that you should expect superstition-like behavior wherever you find animals capable of associative learning, which is nearly everywhere in the animal kingdom.
Monkeys That Invent Rankings From Nothing
Some of the most compelling evidence for genuine animal superstition comes from rhesus monkeys tested on a task that measured whether they impose structure on structureless information. Researchers gave monkeys two sets of pictures. One set had a hidden order: choosing the higher-ranked image earned a reward, and the monkeys could learn the ranking through trial and error. The other set had no order at all; feedback was completely random, with rewards delivered regardless of which image the monkey chose.
The monkeys learned the ordered set just fine. But the striking finding was what happened with the unordered set. The monkeys behaved as if these random images were ordered too, developing consistent preferences that transferred to novel untrained pairs of images. They acted as though they had discovered a ranking that did not exist.5PubMed Central. Superstitious learning of abstract order from random reinforcement
The researchers made the test harder by switching to a reward schedule that actually punished preferences: whichever image the monkey picked less often received more rewards. Even under these conditions, the monkeys maintained their invented rankings. They persisted in their superstitious ordering even when it cost them food. A standard reinforcement-learning computer algorithm could replicate part of the effect under normal conditions but failed under the preference-punishing schedule, suggesting the monkeys were using something more sophisticated than simple stimulus-reward associations. The researchers interpreted this as evidence that the monkeys were inferring abstract structures from objectively random events, using cognitive shortcuts that go beyond basic conditioning.5PubMed Central. Superstitious learning of abstract order from random reinforcement
This is about as close to genuine superstition as has been demonstrated in a non-human animal. The monkeys were not simply repeating a body movement that happened to precede food; they were building a mental model of an ordering system that did not exist and applying it to situations they had never encountered.
Great Apes and Ritual-Like Behavior
Chimpanzees, our closest relatives, show behaviors in the wild that look eerily like proto-rituals. Researchers examining spontaneous ritualistic behavior in non-human great apes identified at least three candidate cases. One is the famous “rain dance” some chimpanzee populations perform at the onset of heavy rain, in which males charge, sway, and display dramatically. Another is a behavior observed in some communities where chimpanzees insert a piece of grass or straw into their ear and leave it there, a behavior that appears to spread socially within groups despite serving no clear practical purpose. A third involves surplus nest-building procedures: constructing nests with more care or elaboration than needed for their functional purpose.6PubMed Central. Spontaneous (minimal) ritual in non-human great apes?
At least two of these three cases revolve around outcomes, meaning the apes appear to associate the behavior with something happening (or not happening), which is the basic structure of superstition. The rain dance is particularly interesting because it occurs reliably at the start of storms. The chimps are not just excited or agitated; the displays have a structured, repeated quality that suggests something beyond a simple startle response. Whether this constitutes genuine superstition in the human sense or a more basic form of ritualized response to environmental triggers is still debated, but it sits in a gray zone that is difficult to explain without invoking some form of false causal association.
The grass-in-ear behavior is a different kind of puzzle. It has no apparent connection to any outcome at all, yet it spreads through social learning within groups. This makes it look more like an arbitrary cultural convention than a superstition, but the boundary between the two is fuzzy. In human societies, many superstitions are also arbitrary cultural conventions that spread socially.
Rats and the Brain’s Built-In Filter
Rats have provided some of the clearest evidence for a neural mechanism that either enables or prevents superstitious behavior. In a classic experiment, rats with damage to the hippocampus or septum were placed in an operant chamber where food pellets arrived on a fixed time schedule, completely independent of anything the rat did. Normal rats and rats with sham (fake) surgeries quickly figured out that their actions did not control the food and stopped pressing the bar unnecessarily. But rats with hippocampal or septal lesions kept pressing, behaving as though their bar-pressing controlled the food delivery.7PubMed. Superstitious bar pressing in hippocampal and septal rats
The key detail is that this was not simply an inability to stop doing things. When the food was discontinued entirely, the lesioned rats stopped pressing quickly. They could inhibit their behavior just fine when no food was arriving at all. What they could not do was recognize that the food was arriving independently of their actions. The finding suggests the hippocampus normally integrates information about the relationship between actions and outcomes, acting as a filter that lets healthy brains eliminate superfluous behaviors. Damage the filter, and the animal defaults to superstitious responding.
This has implications beyond rats. If the hippocampus serves this kind of reality-checking function across mammals, then variations in hippocampal function from one individual to another, or from one species to another, could help explain why some animals are more prone to superstitious behavior than others.
Dopamine and the Urge to Find Meaning
The neurotransmitter dopamine appears to play a central role in the brain’s pattern-detection machinery, and by extension, in superstitious tendencies. One hypothesis proposes that random fluctuations in dopamine firing in reward pathways can produce the same kind of false association that non-contingent reinforcement creates in a conditioning experiment. When dopamine signals arrive unpredictably, the brain may treat them as evidence that something meaningful has happened, even when nothing has.8PubMed. Delusions, superstitious conditioning and chaotic dopamine neurodynamics
Research in humans found that people who endorse paranormal beliefs tend to have a lower threshold for saying “yes, I see a pattern,” favoring false positives over false negatives in perceptual tasks. When these individuals were given a drug that boosts dopamine, their detection threshold actually became more conservative, moving closer to skeptics’ baseline. Skeptics given the same drug shifted slightly in the other direction.9Journal of Cognitive Neuroscience. Dopamine, Paranormal Belief, and the Detection of Meaningful Stimuli The implication is that dopamine levels modulate where the brain sets its threshold for pattern detection. Since dopamine systems are broadly conserved across vertebrates, similar mechanisms likely influence how readily other mammals and birds form false causal associations.
Pigeons Under More Controlled Conditions
Despite the complications with Skinner’s original setup, pigeons have continued to be useful for studying how superstitious responding actually works. In more controlled experiments, pigeons pecking keys under schedules that mixed real reinforcement with periods of no reinforcement showed that superstitious responding could persist for dozens of sessions. When one response key delivered food on a variable schedule and another key was on extinction (never delivering food), pigeons sometimes continued pecking the extinct key for 30 to 45 sessions before the behavior died out. The persistence was greater when the overall rate of food delivery from the other key was higher.10Behavioural Processes. Superstitious responding and reinforcement rate under concurrent variable-interval extinction schedules
This makes intuitive sense: in an environment where good things are happening frequently, it is harder for the brain to figure out which specific action is causing them and which is irrelevant. A pigeon in a food-rich environment has more reason to keep trying everything that might be working. The finding connects to the broader evolutionary point that superstitious associations are more likely to form and persist when the perceived benefits are high relative to the cost of maintaining the behavior.
When Superstition Costs Little, It Sticks Around
The models of superstition evolution make a specific prediction about when superstitious behavior should be most common: when the cost of being wrong is low relative to the perceived benefit of being right. In formal terms, superstitions are more likely when the action itself is cheap and the potential payoff seems large. They are also more sensitive to prior expectations: an animal (or person) that already has some reason to believe a relationship exists will be more likely to keep “confirming” it through coincidence.11Animal Behaviour. The evolution of superstition through optimal use of incomplete information
This helps explain why certain contexts seem to breed superstitious behavior. A pigeon in a lab box has nothing else to do, so the cost of maintaining a useless behavior is nearly zero. A monkey in a testing booth that keeps picking the same image from a random set pays very little for the wrong choice. A chimpanzee performing a rain display at the start of a storm wastes some energy but risks nothing serious. In all these cases, the cost-benefit math favors maintaining the false belief.
The flipside is also informative. In high-cost environments, animals should shed superstitious behaviors more quickly because acting on false beliefs becomes expensive. This prediction aligns with the observation that superstitious key-pecking in pigeons fades faster when the overall reinforcement rate is low: when food is scarce, the bird cannot afford to waste effort on actions that do not produce anything.
Crows and Social Transmission of Associations
One question that comes up when discussing animal superstition is whether false associations can spread socially, turning an individual’s mistake into a group behavior. Crows provide a relevant case study, even though the behavior in question involves a real threat rather than a false one. When researchers in Seattle trapped, banded, and released wild crows while wearing a distinctive mask, the captured birds learned to scold anyone wearing that mask. That individual learning is straightforward. But crows that were present during the trapping scene but were never captured themselves also learned to scold the mask, apparently through conditioning to the mobbing response. And later, lone crows that had never been anywhere near the trapping event also scolded the mask, suggesting horizontal social learning had spread the association through the community.12PubMed Central. Social learning spreads knowledge about dangerous humans among American crows
In this case the association happened to be correct: the masked person really had been dangerous. But the mechanism, social transmission of a learned association about a specific stimulus, is exactly the kind of channel through which a false association could spread just as easily. If a crow had initially mob-called at someone wearing a different, harmless mask and other crows joined in, the same social learning pathway could propagate and sustain an inaccurate threat assessment across the group. The machinery for cultural superstition is there, even if the documented examples involve legitimate threats.
How Humans Compare
One thing that separates human superstition from the animal version is language. Humans can tell each other about their superstitions, codify them into rules (“never open an umbrella indoors”), and pass them across generations as verbal traditions. Animal superstitions, by contrast, are mostly locked inside individual learning histories or spread through direct behavioral observation. A rat cannot tell another rat about its lucky bar-pressing technique.
There is also evidence that human superstitious behavior can interfere with more efficient learning strategies. When people develop rigid response patterns based on coincidental reinforcement, those patterns can prevent them from discovering the actual rules governing a task.13The Psychological Record. Superstitious Behavior and Response Stereotypy Prevent the Emergence of Efficient Rule-Governed Behavior in Humans In other words, superstition does not just add a harmless extra behavior; it can actively block learning. This is a cost that human superstition shares with animal superstition, but the stakes are arguably higher for humans because we rely so heavily on rule-based reasoning.
The dopamine findings described earlier suggest that the neurobiological machinery underlying superstition is conserved across species, with the threshold for pattern detection varying between individuals based on factors like dopamine levels and receptor sensitivity. The human version adds layers of cultural reinforcement, narrative explanation, and deliberate transmission, but the underlying engine is recognizably the same one running in a pigeon, a rat, or a monkey.
The Species Roster
Pulling the evidence together, here is where different animal groups stand in terms of demonstrated superstition-like behavior:
- Pigeons: Show persistent responding to stimuli and actions unrelated to food delivery, though much of the classic evidence likely reflects species-typical foraging rather than genuine false causal learning.
- Rats: Normal rats filter out spurious action-outcome associations efficiently, but damage to the hippocampus produces clear superstitious responding, demonstrating that the tendency exists and is actively suppressed in intact animals.
- Rhesus monkeys: Infer abstract orderings from entirely random feedback and maintain these invented structures even when they reduce reward rates, the strongest laboratory demonstration of cognitive superstition in a non-human animal.
- Great apes: Display proto-ritualistic behaviors in the wild, including rain displays and socially transmitted arbitrary habits, at least some of which appear to revolve around perceived outcomes.
- Crows: Possess social learning mechanisms that can transmit threat associations through populations, providing a channel through which false associations could spread culturally.
The evolutionary models predict that any organism capable of associative learning should show superstition-like tendencies, and the laboratory evidence broadly supports this. The real variation is not in whether a species can be superstitious but in how elaborate and persistent the false belief becomes, which tracks roughly with cognitive complexity. A pigeon’s superstition is a repeated motor pattern. A monkey’s superstition is an invented mental ranking system. A chimpanzee’s superstition can become a group ritual. The gradient from simple to complex mirrors the gradient of cognition itself.
Fish, Insects, and the Boundaries of Superstition
An obvious question is how far down the complexity ladder superstition-like behavior extends. The evolutionary models do not specify a cutoff; they predict that any organism trading off false positives against false negatives in pattern detection should show the effect. Fish are capable of associative learning and conditioning, so in principle they could form spurious associations. Insects have simpler nervous systems but still learn associations between stimuli and rewards. A honeybee that visits a flower and finds nectar may later return to a structurally similar flower that happens to be empty, persisting for some time before giving up. Whether this counts as superstition depends on how loosely you define the term.
Most researchers reserve “superstitious behavior” for cases where an animal acts on a false causal link between its own action and an outcome, not just overgeneralizing from one stimulus to a similar one. By that standard, the evidence gets thin below birds and mammals. That does not mean simpler organisms are immune to false associations, just that demonstrating genuine causal reasoning (as opposed to simple stimulus generalization) is much harder in animals with less complex brains. The monkey experiment, where animals imposed abstract order on random data, sets a high bar that is difficult to test in a fish or a beetle.
Where researchers have looked in simpler organisms, they tend to find the building blocks of superstition: overgeneralization, persistence of learned responses past the point where they are useful, and sensitivity to coincidental pairings. Whether these fragments assemble into anything deserving the label “superstition” remains an open question, and may ultimately be more about semantics than biology.