Elephants are the largest living land animals, in three species across Africa and Asia. They live in matriarch-led family groups, communicate over long distances including through the ground, and appear to address each other with individually specific calls.
Almost everything remarkable about elephants comes back to two things: they live a long time, and they live together. An old matriarch carries decades of social and geographic knowledge that her whole family depends on, which means poaching removes information as well as animals. They sort humans into categories by voice with enough precision to distinguish one ethnic group from another, one sex from another, adults from boys. A 2024 study suggests they may address each other with something functioning like names. And they respond to their dead in ways that are consistently reported and consistently over-interpreted — the behaviour is real, the mourning is an inference nobody has been able to test.
In-depth record · 92% complete · reviewed 2026-08-09
What this page covers
Three living species. The cognition research is split unevenly between them, and the split matters when reading any "elephants can…" headline.
Often confused with: Mammuthus primigenius (woolly mammoth — extinct, and a separate genus)
Quick facts
Species
Three: African savanna, African forest, Asian
Social structure
Matriarch-led family groups of related females and young
Long-distance calls
Low-frequency rumbles that also travel through the ground
Cancer
Around 20 copies of the TP53 tumour-suppressor gene, against one in humans
African savanna elephants, African forest elephants and Asian elephants were long treated as two species and are now recognised as three. They differ in size, ear shape, tusks and habitat — and, importantly for reading research, in how much has been studied about each.
Species
Where
Most of the research on
African savanna elephant
Sub-Saharan grassland and bushland
Social knowledge, communication, human categorisation
African forest elephant
Congo Basin rainforest
Ecology and conservation; least studied cognitively
Asian elephant
South and Southeast Asia
Mirror self-recognition, tool modification, working animals
One of the clearest cases anywhere of an individual animal holding knowledge a group depends on.
Older elephant matriarchs hold knowledge the whole family depends on
Well supported
Good evidence backs this, though some details remain open.
African savanna elephant families led by older matriarchs discriminated familiar from unfamiliar contact calls more accurately and produced more appropriate defensive responses than families led by younger females.
Who this applies to
African savanna elephants at Amboseli, Kenya
Studied in
Loxodonta africana
Why we rate it this way, and what the caveats are
Well supportedHigh confidence
A controlled playback experiment across 21 family groups, resting on decades of individual life-history records that make matriarch age reliable.
How far it can be extended
All three elephant species live in matriarch-led family groups with long lifespans and extensive social memory, so the mechanism is expected to be general. It has been tested experimentally in this population.
Caveats
Playbacks measure short-term response, standing in for decisions that are hard to observe directly.
Matriarch age cannot be randomised, so the comparison is observational.
Still unanswered
How long does a family group take to recover the knowledge lost when an old matriarch dies?
McComb et al., 2014 · Proceedings of the National Academy of Sciences
Extends the same social-knowledge finding to categorising human threat.
This has a hard practical consequence. Ivory poaching takes the largest animals, which means the oldest ones, which means the ones carrying the information. A population count will not show what was lost.
Words used here
Matriarch
The oldest female in an elephant family group, who leads its movements and decisions.
Playback experiment
Playing a recorded sound to a wild animal from a hidden speaker and recording its response — a way of testing what a sound means to it.
Elephants can tell dangerous groups of people from safer ones by voice and by smell
Well supported
Good evidence backs this, though some details remain open.
African savanna elephant families responded defensively to recordings of adult Maasai male voices, and much less to Kamba male, Maasai female or Maasai boy voices saying the same phrase. The same population distinguished the two groups by garment scent, and reacted to the associated colour with aggression rather than avoidance.
Who this applies to
African savanna elephants at Amboseli, Kenya
Studied in
Loxodonta africana
Why we rate it this way, and what the caveats are
Well supportedHigh confidence
The controls are unusually good: the same sentence, differing only in speaker, played from a concealed speaker to 47 family groups. Two independent field experiments, using different senses, agree — and the fact that scent and colour produce different responses is hard to explain as simple learned aversion to one stimulus.
Caveats
The categories are learned from local history, not innate — elephants elsewhere would have learned different ones.
This is categorisation by group, not recognition of individual people.
A group-level response can be driven by one reactive animal.
Still unanswered
Can elephants recognise specific individual humans, as crows demonstrably do?
How does an elephant too young to have experienced conflict acquire the category?
An independent test in the same population using a different sense: garment scent produced avoidance, while the associated colour alone produced aggression.
How we know
Playing human voices to wild elephants
Do elephants sort humans into categories based on how dangerous they are?
In Amboseli, Kenya, elephants have a long history of conflict with Maasai men, who traditionally herd cattle and have speared elephants, and a much more peaceful history with Kamba men, who mostly farm. Researchers recorded men from both groups saying the same phrase in their own language, along with Maasai women and boys. They then played the recordings to elephant families from a hidden speaker and filmed what happened.
What happened
Families reacted defensively to adult Maasai male voices — bunching together, retreating, raising trunks to smell the air. The same phrase from a Kamba man, or from a Maasai woman or boy, produced far less reaction.
What it shows
Elephants categorise humans by voice along lines that track real risk, distinguishing not just language group but sex and rough age within it.
What it does not show
It does not show that elephants recognise individual people by voice, and the categories are learned from local history rather than innate. An elephant elsewhere in Africa would have learned something different.
The controls — what makes this evidence rather than a story
Every speaker said the same sentence, so the words were not the variable.
Women and boys from the same community tested whether the response was to the group or to adult males specifically.
Playbacks came from concealed speakers, so no human was visible.
Whether elephants are dangerous has an uncomfortable answer: yes, and the reason is mostly us. Elephants kill several hundred people a year, overwhelmingly in places where crops, settlements and elephant range have grown into each other. Most incidents involve a crop raid, a startled animal, a cow with a calf, or a bull in musth — a periodic state of very high testosterone in which behaviour becomes genuinely unpredictable. An elephant that has been harassed or injured by people is more dangerous than one that has not, and populations subject to poaching behave measurably differently towards humans.
Adults have no natural predators, which is unusual enough to shape the species: lions and hyenas take calves occasionally and a healthy adult is effectively unkillable by anything except people. That absence is part of why elephant societies can be as slow, long-lived and knowledge-dependent as they are — there is time to accumulate what a matriarch knows.
Safety
Around elephants
Keep distance, stay in a vehicle where one is provided, and never position yourself between an elephant and its calf or its route to water. Treat a bull showing musth signs — temporal gland secretion, dribbling urine — as a reason to leave the area entirely. Local guides and park rules are the authority.
Where this applies: Approach rules and permitted activity are set by individual parks and national authorities and vary widely.
When to get help: Follow the guidance of local rangers or licensed guides, who know the individual animals.
Elephants may address each other with something like names
Emerging evidence
Real findings exist, but too few or too recent to be settled.
Acoustic analysis of wild African savanna elephant rumbles predicted the intended receiver above chance, and playback subjects responded more strongly to calls originally addressed to them — with labels that do not appear to be imitations of the receiver's own call.
Who this applies to
African savanna elephants in two Kenyan populations
Studied in
Loxodonta africana
You may have heard
“Elephants have names for each other, just like humans.”
The comparison to human names rests on one feature: unlike dolphins and parrots, which copy each other's signature calls, elephants appear to use a label that is not an imitation. That is a genuinely striking result and it is one recent study of two populations, with the acoustic analysis carrying real uncertainty.
Why we rate it this way, and what the caveats are
Emerging evidenceModerate confidence
A carefully designed recent study combining machine learning with playback tests. Model accuracy was above chance but modest, the playback sample was small, and independent replication has not yet appeared.
Caveats
Above-chance prediction is a long way from reliable decoding.
Two populations in Kenya; untested elsewhere.
Recent enough that replication is still outstanding.
Still unanswered
Are the labels learned, and do they persist over an animal's lifetime?
The intended receiver was predictable from call structure, and elephants reacted more strongly to calls addressed to them.
Elephants pick up signals travelling through the ground
Well supported
Good evidence backs this, though some details remain open.
Wild African savanna elephant herds responded with alerting and defensive bunching to alarm calls transmitted purely as ground vibration, with no airborne sound component.
Who this applies to
African savanna elephants
Studied in
Loxodonta africana
Why we rate it this way, and what the caveats are
Well supportedModerate confidence
A direct field playback with the airborne channel removed. Moderate confidence because outdoor playbacks are hard to isolate completely and the detection pathway is still being worked out.
How far it can be extended
The anatomy involved — large fat pads in the feet and a heavy skull suited to bone conduction — is shared across elephant species, though playback tests have been done on African savanna elephants.
Caveats
"Elephants hear with their feet" is a shorthand — the detection route, through feet, trunk and bone conduction, is not fully resolved.
Limited replication.
Still unanswered
How far do seismic signals usefully travel in real habitat?
O'Connell-Rodwell et al., 2006 · Behavioral Ecology and Sociobiology
Herds responded to seismic-only transmission of alarm calls.
Unlike dolphins and parrots, which address each other by copying the other animal's signature call, elephants appear to use a label that is not an imitation — which is what makes the comparison with human names interesting rather than loose.
Elephants can learn to produce sounds they have heard, including ones no elephant makes
Well supported
Good evidence backs this, though some details remain open.
African elephants have been documented producing vocalisations outside the species repertoire that closely match sounds in their acoustic environment — motorway traffic in one orphaned female, Asian elephant chirps in a male raised among them — demonstrating vocal production learning.
Who this applies to
African elephants; two documented individuals
Studied in
Loxodonta africana
Why we rate it this way, and what the caveats are
Well supportedHigh confidence
Vocal production learning is one of the rare capacities that a single unambiguous case demonstrates: an animal producing a sound its species does not have cannot have inherited it. Two independent cases with acoustic analysis.
How far it can be extended
Vocal learning has since been reported in Asian elephants as well, and the capacity is a species-level property that a single unambiguous case can establish.
Caveats
Two individuals in unusual circumstances; how often wild elephants learn calls is unknown.
What function the capacity serves naturally is not addressed by these cases.
Acoustic similarity is measured; the route of imitation is inferred.
Still unanswered
Do wild elephants use vocal learning to converge on family-specific calls?
How is a trunk-and-larynx vocal tract capable of this range?
Individually specific calls in wild elephants — a capacity that requires vocal learning as a precondition.
Two individuals is normally a fatal weakness in a claim about a species, and here it is not, because of what is being claimed. Vocal production learning is demonstrated by a single unambiguous case: an elephant producing a sound elephants do not make cannot have inherited it. One orphaned female housed near a motorway produced truck-like sounds; a male raised among Asian elephants produced their chirps.
The capacity matters because of what it makes possible. An elephant cannot address another by a learned, individually specific call unless elephants can learn to make new sounds at all — so this result is the precondition for the name-like calls documented in wild elephants nearly twenty years later.
Words used here
Vocal production learning
Modifying the sounds you make on the basis of what you have heard. Rare in mammals — humans, cetaceans, bats, seals and elephants.
The study that gives the matriarch literature its stakes.
Families led by older matriarchs lose fewer calves in a severe drought
Well supported
Good evidence backs this, though some details remain open.
During the 1993 Tarangire drought, elephant family groups led by older matriarchs suffered substantially lower calf mortality and were more likely to leave the park for areas holding water — a pattern consistent with those matriarchs having experienced a comparable drought decades earlier.
Who this applies to
African elephant families in Tarangire, Tanzania, during one severe drought
Studied in
Loxodonta africana
Why we rate it this way, and what the caveats are
Well supportedModerate confidence
A clear survival difference tracking matriarch age, with a plausible mechanism supported by independent work on matriarch decision-making. Confidence is moderate because a single event cannot separate memory from other properties that correlate with a matriarch’s age.
Caveats
One drought; matriarch age correlates with group size and range familiarity as well as with memory.
The memory explanation is inferred from age rather than demonstrated.
No comparable manipulation is possible or ethical.
Still unanswered
How is ecological knowledge held and used — as remembered places, remembered routes, or learned rules?
How long does a family take to recover after losing an old matriarch?
Douglas-Hamilton et al., 2006 · Applied Animal Behaviour Science
Context on the social centrality of matriarchs within a family.
Earlier work showed that older matriarchs respond more appropriately to unfamiliar contact calls. This shows the consequence in the only currency that counts: during a severe drought in Tarangire, families led by older females lost fewer calves, and were more likely to leave the park for water. A drought of comparable severity had occurred within those matriarchs’ lifetimes, decades earlier.
The conservation implication is hard and NatureHQ states it rather than softening it. Ivory poaching removes the oldest animals preferentially, because they carry the largest tusks. A population can recover its numbers long before it recovers the knowledge needed to survive the next rare event.
At least one Asian elephant has passed the mirror mark test
Emerging evidence
Real findings exist, but too few or too recent to be settled.
Of three Asian elephants tested with a mirror and a visible head mark plus an invisible control mark, one repeatedly touched the visible mark while viewing herself and did not touch the control.
Who this applies to
Asian elephants — three individuals, one of which passed
Studied in
Elephas maximus
You may have heard
“Elephants recognise themselves in mirrors.”
One elephant out of the three tested touched the mark. That is a real result and worth reporting; stating it as a fact about elephants in general skips the number that makes it interpretable.
Why we rate it this way, and what the caveats are
Emerging evidenceModerate confidence
A single unambiguous pass with a proper invisible-mark control is meaningful, and pass rates are low even in chimpanzees. One animal out of three is still one animal out of three.
Caveats
One of three animals passed.
The mark test is visual, and elephants lead with smell and touch — a failure may say more about the test than the animal.
Captive animals with extensive human contact.
Still unanswered
Would a test built around smell or touch give different results?
Plotnik et al., 2006 · Proceedings of the National Academy of Sciences
One of three elephants passed the mark test; all three used the mirror non-socially.
Elephants shorten branches to make better fly swats
Well supported
Good evidence backs this, though some details remain open.
Working Asian elephants observed in Nepal modified branches — shortening them and stripping side shoots — before using them to switch flies, with modified branches associated with more effective fly control.
Who this applies to
Asian elephants, in this study working animals in Nepal
Studied in
Elephas maximus
Why we rate it this way, and what the caveats are
Well supportedModerate confidence
Systematic observation with quantified modification and effectiveness. Moderate confidence because the animals lived in constant contact with people, so learning from humans cannot be excluded.
How far it can be extended
Branch use for fly switching is widely reported in both African and Asian elephants; the systematic evidence for modification comes from this Asian elephant study.
Caveats
Working animals with lifelong human contact.
Thirteen individuals.
Effectiveness measures are observational rather than experimental.
Still unanswered
Do wild elephants modify branches to the same degree?
Elephants modified branches to a more effective size and shape before use.
The mirror test is worth a note of caution. It is a visual test, applied to an animal whose primary senses are smell and touch. A failure may say as much about the test as about the animal.
Elephants show strong, repeated interest in dead elephants — whether that is grief is unknown
Emerging evidence
Real findings exist, but too few or too recent to be settled.
African savanna elephants approach, touch, smell and attempt to lift dying and dead conspecifics, including unrelated individuals, and revisit remains. No study design to date can determine the internal state involved.
Who this applies to
African savanna elephants
Studied in
Loxodonta africana
You may have heard
“Elephants mourn their dead and hold funerals.”
The behaviour — approaching, touching, returning to remains — is real and repeatedly observed. "Mourning" attributes an inner experience that no field observation can establish or rule out. NatureHQ reports the behaviour confidently and the interpretation cautiously, because that is where the evidence sits.
Why we rate it this way, and what the caveats are
Emerging evidenceModerate confidence
The behaviour itself is widely and independently reported, so we are confident it occurs. The published detailed accounts are largely observational and often single events, and the emotional interpretation is untestable with current methods.
How far it can be extended
Similar behaviour towards the dead has been reported in Asian elephants and across many African sites by independent observers, though systematic study remains limited.
Caveats
Much of the published record consists of individual observed events without controls.
Interest in a dead conspecific has several possible explanations, from social bond disruption to plain curiosity.
Observer presence may influence what is seen.
Still unanswered
Do elephants respond differently to a dead individual they knew well?
Is there any measurable physiological correlate of the response?
Douglas-Hamilton et al., 2006 · Applied Animal Behaviour Science
A detailed record of individuals from five families, including unrelated animals, attending a dying matriarch.
This is the most emotionally loaded subject on the page, and the one where the gap between the behaviour and the interpretation is widest. NatureHQ states the behaviour confidently, because it is repeatedly and independently observed, and states the interpretation cautiously, because no method available can settle it.
Elephants get far less cancer than their size and lifespan would predict
Well supported
Good evidence backs this, though some details remain open.
Estimated cancer mortality in captive elephants is low relative to body mass and lifespan. Elephant genomes carry around twenty copies of the tumour-suppressor gene TP53 against one in humans, and elephant cells show elevated programmed cell death after DNA damage.
Who this applies to
African and Asian elephants
Studied in
Loxodonta africana, Elephas maximus
Why we rate it this way, and what the caveats are
Well supportedModerate confidence
The gene-copy finding is robust and easily replicated. The causal link to low cancer rates is supported by cell experiments but not demonstrated in living animals, and the mortality estimates come from captive records with known biases.
How far it can be extended
TP53 copy number was examined across elephant species and compared with other mammals.
Caveats
Cancer rates estimated from zoo records, which are incomplete.
Cell-line behaviour is a long way from whole-animal cancer resistance.
Other mechanisms, including a reactivated pseudogene reported by a separate group, may contribute.
Still unanswered
How much of the resistance is attributable to TP53 copy number specifically?
Does this suggest anything usable for human medicine, or is it particular to elephants?
Documented the extra TP53 copies and the heightened apoptotic response of elephant cells.
The general puzzle is called Peto's paradox: a large, long-lived animal has vastly more cells dividing over a lifetime than a mouse, so it ought to develop far more cancer. It does not. Elephants are the best-studied attempt at an answer.
Words used here
TP53
A gene that helps detect damaged DNA and can trigger a cell to destroy itself rather than risk becoming a tumour.
An elephant is pregnant for about 22 months — the longest gestation of any land mammal — and produces a single calf weighing around 100 kilograms. The length is not arbitrary: a calf must arrive with a brain developed enough to keep up with a moving herd and to begin learning from it immediately, and elephant brain development is unusually protracted.
Gestation around 22 months; twins occur and are rare.
Calves nurse for two to four years and stay with their mother far longer, into adulthood for females.
Females remain in the natal herd for life; males leave in adolescence and live alone or in loose bachelor groups.
Females typically have a calf every four to five years, which makes population recovery from losses very slow.
Wild lifespan is commonly 50 to 70 years, and captive lifespans are frequently shorter rather than longer.
The last point is unusual among large mammals and worth stating plainly, since the assumption normally runs the other way. Zoo elephants have historically had shorter median lifespans than wild populations, which is one of the reasons elephant keeping has changed substantially and why some collections have stopped.
Teeth set the ceiling on elephant life. An elephant has six sets of molars in a lifetime, and rather than erupting from below they move forward along the jaw, each worn set pushed out by the next. When the sixth is worn away — usually in the sixth or seventh decade — the animal can no longer process enough vegetation, and it starves. Old age in an elephant is a dental event.
A calf uses its trunk badly for months. It has around 40,000 muscle units and no bones, and learning to control it is a skill acquired over a year or more — young calves frequently step on their own trunks or drink with their mouths instead.
Words used here
Natal herd
The family group an elephant is born into. Females stay in it for life; males leave.
Molar progression
Elephant molars move forward along the jaw and are replaced six times. When the last is worn out, the animal cannot feed.
The trunk is a fused nose and upper lip, containing no bone and no joint, and estimated to hold on the order of 40,000 muscle fascicles. Having no skeleton means it can bend anywhere, and having no joint means it moves by contracting muscle against the incompressible volume of its own tissue — the same principle as an octopus arm or a human tongue.
What it does is a long list, and the list is the point: breathing, smelling, drinking, feeding, dusting, greeting, touching, disciplining calves, producing sound, and manipulating objects with the fingerlike tip — one in Asian elephants, two in African. It is simultaneously the animal’s nose, its hand and a large part of its social apparatus.
Drinking is the one most often got wrong. An elephant does not drink through its trunk in the way a straw is used; it draws water up into the trunk — around eight litres — and then squirts it into its mouth. Water going all the way through would arrive in the lungs.
Elephants swim well and use the trunk as a snorkel, which is possible because the nostrils run its full length. It is the only mammal that can breathe at the surface while walking on the bottom.
Words used here
Muscular hydrostat
A structure that moves by contracting muscle against its own fluid volume rather than against a skeleton. A trunk, an octopus arm and a tongue all work this way.
An elephant is a herbivore that is bad at digesting plants, and almost everything about how it lives follows from that. A hindgut fermenter extracts perhaps 40–50% of what it swallows, against the 70% or more a cow gets from the same material by fermenting it in a rumen first and chewing it twice. The elephant’s answer is not better digestion but more throughput: 150–200 kg of vegetation a day, taken over 16 to 18 hours, with the remainder passing through largely intact.
That inefficiency is ecologically useful in a way efficiency would not be. Undigested seeds pass through and are deposited in dung, and for several tree species elephants are the principal — occasionally the only — long-distance disperser. An animal that digested its food properly would destroy the seeds.
Grass, leaves, bark, roots, fruit and stems, in proportions that shift with season and species. Savanna elephants graze more in the wet season and browse more in the dry.
Bark is stripped with the tusks in lean months, which is why elephants kill trees — the tree is not being eaten so much as opened.
Around 100–200 litres of water a day, drunk by squirting rather than sucking.
Salt and minerals are dug for; some populations mine caves for sodium and have done so for generations.
Six sets of molars arrive across a lifetime, moving forward like a conveyor. When the last set wears out the animal starves, and that is what old age looks like in an elephant.
Elephants are not omnivores and do not hunt. Reports of meat-eating are rare, opportunistic and disputed, and nothing in the dentition or gut supports it as part of the diet. The peanut association, meanwhile, has no basis at all in the wild — peanuts grow underground in South America, elephants live in Africa and Asia, and the pairing comes from circus feeding rather than from biology. Captive elephants will eat them, as they will eat most things offered.
A tooth is what usually ends the life. Molars are replaced six times and no more, so a wild elephant’s lifespan is set less by disease or predation than by the point at which it can no longer grind its food.
Words used here
Hindgut fermenter
An animal that ferments plant material after the stomach rather than before it. Faster and less thorough than a ruminant.
Browse and graze
Browsing is eating leaves, twigs and bark; grazing is eating grass. Elephants do both, in shifting proportion.
Most of the body answers one question: how to shed heat at six tonnes.
Large animals overheat easily, because heat is produced in proportion to volume and lost in proportion to surface area, and volume grows faster. An elephant also has almost no sweat glands. The ears are the principal solution: an African elephant’s ear is a thin sheet carrying a dense network of vessels close to the surface, and flapping it moves blood through that radiator and air across it. Blood returning from the ears can be several degrees cooler than blood entering them.
The size difference between the species follows the climate rather than the family tree. African savanna elephants live in hotter, more open country and have far larger ears than Asian elephants, which evolved in shadier forest. The skin helps too: its deep wrinkles hold many times more water than smooth skin would, so a bath keeps cooling for hours after it ends.
Throwing dust and mud is part of the same system rather than play. A coating of mud is sunscreen, insect repellent and — as it dries and is rubbed off against a tree — a way of removing parasites. Elephants that cannot reach mud will throw dry dust instead, which works less well and is done anyway.
An elephant cannot jump, and is the only land mammal of its size that cannot. The legs are pillars built for weight-bearing, and all four feet never leave the ground at once — which is also why an elephant has no true run.
Top speed is around 25 km/h in a fast walk, sustained only briefly. It looks like running and is mechanically not.
Elephants lie down to sleep, usually for a couple of hours, and sleep standing as well. A very large elephant lying down for too long risks damaging its own organs under its weight.
Feet spread under load and contract when lifted, and the fatty pad in the sole transmits ground vibration to the bones — part of how elephants detect seismic signals from kilometres away.
Swaying is the one behaviour on this page that is not an adaptation. Repetitive rocking is seen almost exclusively in captivity, is classed as a stereotypy, and is associated with restricted space and lack of stimulation. It is not something wild elephants do, and reading it as contentment gets it precisely backwards.
The brain weighs around 5 kg — the largest of any land animal, and roughly four times a human’s. The idea that it is the size of a peanut is not a misremembered statistic; it appears to be a confusion with the peanut story, and it is wrong by four orders of magnitude.
Words used here
Stereotypy
A repetitive behaviour with no obvious function, associated with captivity and restricted environments. A welfare indicator, not a mood.
Countercurrent cooling
Running warm blood close to a cooling surface so heat leaves before the blood returns to the core.
Matriarchs shown to hold knowledge the group needs
Playback experiments show families with older matriarchs discriminating familiar from unfamiliar calls more accurately — the first demonstration that an individual animal can be a repository of information for a group.
Identical sentences spoken by different groups produce systematically different defensive responses in 47 family groups.
Changes how the 2007 result reads
The 2007 result could be explained by learned aversion to a smell. Reproducing the categorisation through an entirely different sense, with the sentence held constant, makes a single-stimulus account much harder to sustain.
Elephants appear to address one another with calls specific to the individual, without imitating that individual’s own call — the feature that distinguishes a name from a copy.
The shape of this history is worth noticing. Almost every landmark is a *control* rather than a discovery: the interesting work was not finding that elephants respond to people, which anyone could see, but building designs that separate categorisation from smell, and identity from imitation. The 2024 name-call result is the newest and by some distance the least settled.
Are the name-like calls learned, and do they last a lifetime?
Why it matters: A stable, learned, arbitrary label for an individual would be the closest thing to a name found outside humans.
What would settle it: Longitudinal recording of known individuals across years, and independent replication elsewhere.
What does an elephant understand about a dead elephant?
Why it matters: The behaviour is unmistakable and the interpretation is contested. It is one of the clearest cases where the honest answer is that we do not know.
How long does a family group take to recover the knowledge lost with an old matriarch?
Why it matters: It converts a conservation cost that is currently invisible into something measurable.