There is an intelligence ranking
Intelligence is not one variable measured by a tool puzzle. Scientists compare cognitive domains; they do not award a cross-species crown.
Bird Smart · cognition
Not “are parrots intelligent?” but “which operation, under which design?”
Comparative cognition is full of tiny samples, unusually experienced animals tested repeatedly, and results that depend on the apparatus. Parrots have still produced several genuinely important findings. Here they are, with the caveats the headlines usually drop.
Tool use
A cockatoo that was never a tool specialist invented one anyway
Goffin's cockatoos are not habitual tool users in the way New Caledonian crows are. That is precisely what makes the result interesting.
- Proven
Figaro made his own tools, and the finding held up
In 2012 a male Goffin's cockatoo named Figaro spontaneously fashioned usable splinters and inserted them to retrieve food he could not otherwise reach. Because the innovation came from one animal, the original study was compelling but inherently small-N.
Follow-up work strengthened it considerably. In a social-learning experiment with 12 Goffin's cockatoos, birds that watched a demonstrator were more likely to acquire the solution; three males in the demonstrator condition acquired tool use and two manufactured tools. Their methods were not exact motor copies, which points to emulation of the result rather than high-fidelity imitation. Later studies showed manufacture from different materials and flexible tool selection and transport, and field research has since documented tool manufacture and tool sets in wild Goffin's — which largely removes the worry that this was a laboratory curiosity.
Auersperg et al., Current Biology 2012, 10.1016/j.cub.2012.09.002 · Auersperg et al., Proc. R. Soc. B 2014, 10.1098/rspb.2014.0972 - Proven — one individual
A disabled kea built itself a preening tool
Kea are famously exploratory and have succeeded at many object-manipulation and tool tasks. A striking 2021 case documented a kea with a damaged beak spontaneously modifying and using objects as self-care and preening tools. That is a strong observation of innovation in that individual. It is a case study, so it does not support “kea routinely manufacture grooming tools.”
Reasoning and self-control
Real results, very small rooms
Much of the grey-parrot literature rests on a handful of intensively studied birds. That does not make it worthless; it makes species-wide generalisation unsafe.
- Proven at the task level
Inference by exclusion
African greys have solved exclusion problems — inferring that an unselected alternative must be correct once competing possibilities are ruled out. Pepperberg and colleagues reported this under controlled conditions. The defensible reading is evidence for inferential problem solving within those task constraints, not proof of formal human logic. The number of subjects in this tradition is very small.
- Proven, small-N
Object permanence
Greys have been tested across Piaget-inspired object-permanence problems, including complex invisible displacements, and can maintain a representation of an object that has left view. Performance varies with displacement type, training, and individual — so “parrots have exactly the object permanence of a two-year-old child” is a sentence to avoid.
- Proven phenomenon, variable limits
Delayed gratification depends on the procedure
An early experiment tested only three African greys with an accumulating-food procedure. The birds could delay taking an immediately available reward in some circumstances, but performance was modest and strongly individual-dependent. Later token-delay work found much longer waits in some trained greys, including Griffin.
The methodological point matters more than any single number: “maximum seconds waited” cannot be compared across studies that differ in reward structure, training history, visibility, and procedure. There is no species stopwatch ranking.
Vick, Bovet & Anderson, Animal Cognition 2010, 10.1007/s10071-009-0284-2 - Important methodological caveat
The kea statistical-inference study — and its addendum
Bastos and Taylor tested six kea and reported three striking signatures of statistical inference: choices driven by relative rather than merely absolute frequencies, sensitivity to physical sampling constraints, and use of information about a human sampler's apparent bias. The paper concluded that kea can integrate information across domains in ways previously shown most strongly in humans and great apes.
The authors later published an addendum noting that their procedure had not accounted for the possibility that tokens were sometimes visible between the experimenter's fingers, which could have supplied extra information. That does not erase the result, but it materially weakens how confidently it should be presented. This site labels it preliminary — not “parrots understand probability like humans.”
Bastos & Taylor, Nature Communications 11:828, 2020, 10.1038/s41467-020-14695-1 - Proven, eight birds
Greys voluntarily passed tokens to a partner who needed them
Brucks and von Bayern found that African greys transferred exchangeable tokens to a conspecific when the partner could use them to obtain food. Seven of eight greys transferred tokens in the first trial; blue-headed macaws in the comparison condition transferred substantially fewer.
The correct claim is not “African greys are altruistic people with feathers.” The experiment operationalised prosocial helping as transferring an instrumental object without receiving the partner's reward. Motivation, social relationship, prior experience, and species ecology all remain open questions.
Brucks & von Bayern, Current Biology 2020, 10.1016/j.cub.2019.11.030 - Preliminary / mixed
Token economies are learned, not innate
Greys and several other parrots have taken part in exchange, sharing, reciprocity, and prosocial-choice paradigms, with results differing substantially between tasks and species. Tokens are experimentally useful because they separate the instrumental object from the food reward — but subjects need extensive training on token value, so training history itself shapes performance. The evidence supports sophisticated learned economic decisions in some parrots, not a unified avian theory of money.
Parrots, corvids, apes
Matching an ape on one apparatus is task-level convergence
New Caledonian crows are extraordinary because habitual tool manufacture is part of their species-typical ecology. Goffin's cockatoos are compelling in a different way — they innovate flexible tool solutions without being specialised habitual tool makers. Kea excel at exploration and flexible physical problem solving. Great apes have a far larger comparative literature across causal, social, statistical, and tool cognition.
So when a parrot equals an ape on one test, the finding is that both species can solve that problem. It is not a statement about equal global intelligence, and there is no scientifically accepted ladder reading “human > chimp > dolphin > crow > African grey.”
Why “tiny bird brain” is wrong
Small skulls, primate-scale neuron counts
Proven
The strongest biological reason to drop the phrase “bird brain”
Olkowicz and colleagues counted neurons across bird brains and found that parrots and songbirds — including corvids — pack exceptionally high numbers of neurons into their forebrains. Some bird species consequently possess pallial neuron numbers comparable with, or exceeding, those of much larger-brained primates.
The qualifier is essential: the paper did not say a parrot brain is a primate brain. Avian and mammalian brains differ substantially in development and gross organisation. What the finding explains is how a physically small avian brain can hold enormous computational resources without a mammal-sized skull. It is also one reason a “brain size divided by body size” leaderboard is biologically inadequate.
Olkowicz et al., PNAS 113, 7255–7260, 2016, 10.1073/pnas.1517131113Where the story gets stretched
Three claims to stop repeating
Solving a task proves the inner experience
Passing a test does not establish that the animal has the subjective state humans attach to words like “planning,” “generosity,” or “reasoning.”
One clever bird speaks for its species
Figaro, Alex, and Griffin were individuals — several of them exceptionally trained. Their results describe what a parrot can do, not what yours will do.
Take these away: Goffin's cockatoos make their own tools, and wild Goffin's have been documented producing tool sets. African greys have voluntarily passed useful tokens to a grey that needed them — in a study of eight birds. Six kea produced remarkable probability-like choices, then the study received a methodological addendum. Some parrot forebrains contain primate-like neuron numbers. And a parrot matching an ape on one test does not mean the two species have equal overall intelligence.
Continue
Where to read next
Cognition sits alongside the mirror and screen evidence, the records page where neuron counts and lifespan data reappear, and the symbols and buttons guide, where Alex's controlled transfer tests are covered in full. For the cockatoos in this page as companion animals rather than subjects, see the cockatoo species guide.