A male mourning cuttlefish wants to mate with a female. A rival male is watching. The cuttlefish does something that should be impossible for a creature with a 10-month lifespan and no social upbringing: he splits his body display in half. On the side facing the female, he shows courtship coloration — bright, patterned, unmistakably male. On the side facing the rival, he simultaneously displays female patterning — muted, cryptic, a disguise designed to convince the rival that no competition is present. Two contradictory signals, broadcast from one body, targeted at two different audiences at the same time. That’s not camouflage. That’s not instinct in any simple sense. That’s an animal producing a lie calibrated to two different observers with two different perspectives, and executing it in real time with chromatophores instead of words.
The question of whether animals deceive each other is settled — they do, constantly, across hundreds of species. The question that matters is what kind of deception they’re performing, because the answer tells you something fundamental about what’s happening inside their nervous systems. And a May 2025 paper in Trends in Ecology & Evolution by Drerup, Garcia-Pelegrin, Clayton, and colleagues just reframed the entire field by proposing cephalopods — not primates, not corvids — as the ideal model organisms for studying the most cognitively demanding form of deception.
The spectrum
Not all deception is created equal, and the distinctions matter more than the examples. At the bottom of the spectrum, you have deception that requires no cognition at all. Harmless butterflies evolving wing patterns that mimic toxic species is deception — it communicates false information to predators — but nobody claims the butterfly is “lying.” The misinformation is encoded genetically over evolutionary time, not produced by an individual making a decision. A stick insect that looks like a twig is deceiving every bird that passes without doing anything except existing. This is deception without a deceiver.
One step up, you get deception that involves behavioral flexibility but may still be conditioned rather than cognitively strategic. Firefly femmes fatales — females of the genus Photuris that mimic the flash patterns of Photinus females to lure Photinus males close enough to eat them — produce species-specific flash codes that attract prey. The behavior is adaptive, it’s flexible (the predator adjusts flash timing to match different prey species), but it may operate through relatively simple learning mechanisms rather than any representation of what the victim “believes.”
At the top sits tactical deception — deceptive behavior that is flexibly adjusted based on the identity, perspective, or inferred knowledge of the observer. This is the category that implies something approaching theory of mind, the capacity to understand that another individual’s knowledge or perspective differs from your own and to exploit that asymmetry. Tactical deception has been documented primarily in two vertebrate groups: primates and corvids.
The primate evidence
Primates are the best-studied tactical deceivers. Research across 18 species has demonstrated a strong correlation between the frequency of tactical deception and the size of the neocortex — suggesting that the capacity to deceive conspecifics was itself a selection pressure driving brain evolution, a hypothesis known as the Machiavellian intelligence theory. Chimpanzees suppress food calls when dominant individuals are nearby, concealing discoveries rather than sharing them. Subordinate males lead dominant rivals away from hidden food by walking confidently in the wrong direction, then doubling back to retrieve it when the dominant is out of sight. Baboons have been observed using false alarm calls — predator warnings issued when no predator exists — to scatter competitors away from food resources.
The key distinction is context-dependence. A chimpanzee doesn’t suppress every food call — she suppresses them selectively, when a specific dominant individual is present and when the social cost of sharing outweighs the benefit. The behavior varies with audience, which means the animal is tracking who knows what, who can see what, and what the consequences of being detected are. Whether that constitutes genuine “mind-reading” or a sophisticated learned association between behavioral cues and outcomes is the debate that has occupied comparative cognition researchers for decades. The behavior looks like theory of mind. Proving it is theory of mind rather than behaviorally flexible conditioning is extraordinarily difficult, because the observable output is identical.
The corvid evidence
Corvids — jays, ravens, crows — match primates in deceptive sophistication despite being separated by 320 million years of evolution. Nicola Clayton’s lab at Cambridge has produced some of the field’s most striking results. Western scrub-jays that have been observed caching food by another jay will return later, when the observer is absent, and re-cache the food in a new location — but only if the cacher has personal experience of having stolen food from others. Jays that have never stolen don’t re-cache. The implication is that the cacher is projecting its own experience of thievery onto the observer — reasoning, in effect, “I would steal from that cache, so this jay probably will too.”
Ravens observed by Thomas Bugnyar show similar patterns. They monitor the gaze direction of competitors during caching events and adjust their concealment strategies based on whether they believe the competitor has visual access to the cache location. A 2016 study demonstrated that ravens can track whether an observer can see through a peephole — adjusting their caching behavior based on whether the peephole is open or closed, even when no actual observer is present. The researchers argued this showed an understanding of another’s visual perspective independent of behavioral cues, though the interpretation remains contested.
Garcia-Pelegrin’s work at Cambridge has added another dimension: using magic tricks as experimental tools. Jays were shown sleight-of-hand coin vanishes and real transfers. The birds tracked the real transfers accurately but were fooled by the sleights — demonstrating that they form predictions about object permanence and manual actions that can be violated, just as human audiences are fooled by the same techniques. The cognitive architecture that makes you susceptible to a magic trick is the same architecture that allows you to deceive others.
The cephalopod frontier
The 2025 Drerup et al. paper in Trends in Ecology & Evolution argues that cephalopods — octopuses, cuttlefish, and squid — are the ideal organisms for studying tactical deception because they combine two things no other taxon offers at the same scale: an extraordinarily rich behavioral repertoire of naturalistic deception and cognitive abilities sophisticated enough to potentially support flexible, audience-dependent deployment.
The mourning cuttlefish’s split-body display is the poster case, but it’s not the only one. Common cuttlefish flash false eyespots to scare approaching predators — but only to visually oriented predators, not to those that hunt by smell, suggesting the behavior is calibrated to the sensory capabilities of the audience. Giant Australian cuttlefish males that are too small to win fights adopt female coloration and posture to sneak past rival males and access females — a transient, context-dependent mimicry that is abandoned the moment the social environment changes. Female opalescent squid mimic male appearance by flashing a white stripe to deter unwanted mating attempts, deploying the deception only under specific conditions.
The critical question the paper raises is whether these behaviors constitute conditioning — learned responses to specific cue-outcome pairings — or tactical deception, which requires the deceiver to evaluate information about the observer and adapt its strategy based on the observer’s perspective. The distinction matters because cephalopods have nervous systems organized completely differently from vertebrates — 500 million neurons in an octopus, most distributed across peripheral ganglia in the arms rather than concentrated in a central brain. If cephalopods perform tactical deception, it evolved independently from the primate and corvid lineages, through entirely different neural architecture, which would tell us something profound about what kinds of nervous systems can support perspective-taking and flexible social cognition.
What deception tells us about minds
The ability to lie is — paradoxically — one of the strongest indicators of cognitive sophistication. A truthful signal requires only a detection system and a broadcast mechanism. A deceptive signal requires, at minimum, a model of what the receiver expects, an ability to generate a signal that violates reality while matching that expectation, and — in the case of tactical deception — a capacity to adjust the deception based on who’s watching. Every step up the deception spectrum adds a layer of cognitive complexity that brings the deceiver closer to what we’d recognize as a mind.
The convergent evolution of tactical deception in primates, corvids, and potentially cephalopods — three lineages separated by hundreds of millions of years and running on radically different neural hardware — suggests that the capacity for deception isn’t a quirk of primate brains. It’s a solution that evolution converges on whenever social complexity creates enough pressure to make manipulating others’ behavior worth the cognitive investment. The cuttlefish that splits its body display between two audiences and the chimpanzee that leads a rival away from hidden food are solving the same problem with different equipment. The problem is other minds. The equipment is whatever nervous system natural selection had to work with.
We cover deception alongside mirror neurons, dolphin naming, tool use, and 20 other investigations into what animal nervous systems can do across our Animal Culture & Knowledge course — where the question isn’t whether animals have minds but what kind of minds they have, and how we’d know.
