Gorilla

The family tree is settled and is regularly reported as though it were not. Humans and chimpanzees are each other's closest living relatives; gorillas branched off earlier. Across roughly 15% of the genome the local relationship disagrees with that — a human or chimpanzee stretch may be closer to gorilla than to the other — and this figure gets reported as uncertainty about the tree. It is not. It reflects that a species split is a process rather than an event: for a long period the ancestral population carried variation that then sorted at random into the descendant lineages, and different bits of the genome record different draws.

The framing correction matters more than the phylogeny. Gorillas are not a stage we passed through. They are a sister lineage that has been evolving for precisely as long as the human lineage has, and knuckle-walking — the trait that makes them look like a halfway house — appears from wrist anatomy to have been invented separately by gorillas and by chimpanzees, rather than inherited from a shared knuckle-walking ancestor we then grew out of. That conclusion is contested; the framing correction does not depend on it.

Two species, four subspecies, and the numbers are lopsided in a way that shapes everything people believe about gorillas. Western lowland gorillas number in the hundreds of thousands and are the least observed. Mountain gorillas number around a thousand and are the most intensively studied animals in the genus by a wide margin — habituated, monitored daily, and the source of nearly every image and nearly every generalisation.

The multipliers all trace to the same place: a small number of trials in the 1920s in which captive apes were induced to pull on a dynamometer. The problem with those is not sloppiness but unanswerability. An ape that pulls hard is strong; an ape that does not may be weak, may be bored, may be holding the handle differently. Motivation cannot be controlled in a chimpanzee, and the resulting figures have been repeated, rounded and inflated ever since — four times, six, eight, ten, depending on where you read it.

The way to answer it is to go somewhere motivation does not exist. Individual muscle fibres from chimpanzees were tested directly for force, shortening velocity and power, and their fibre-type composition determined, then scaled to whole-muscle performance against human values obtained the same way. The difference is roughly 1.35-fold, and it comes almost entirely from chimpanzees carrying a higher proportion of fast-twitch fibres. The contractile machinery itself is the same. Ape muscle is ordinary muscle.

Two honest qualifications. Those are chimpanzees — no comparable measurement exists for gorillas and may never be obtainable — so extending it is an argument from shared muscle physiology rather than a measurement. And absolute force is genuinely much greater in a gorilla, because an adult male weighs two to three times what a person does. That is a size difference. Calling it a strength multiplier is a category error, and it is the error the whole mythology is built on.

What gorillas do instead of being ten times stronger is considerably better. Recording chest beats from identified wild silverbacks and relating them to body size measured independently — photogrammetrically, so the two numbers cannot be circular — showed that larger males produce lower-frequency beats, and reliably enough that size is recoverable from the sound.

That makes the display an honest signal in the technical sense: peak frequency is set by the dimensions of the resonating air sacs and body, so a small male physically cannot fake a large male's signature. The honesty is enforced by acoustics rather than by restraint, which is exactly what allows a contest to be settled without a fight. Two males who can each hear how big the other is have far less left to argue about — which is the job the chest-beating gorilla of popular imagination is supposed to be doing, and is not.

Start with what gorillas do without being taught anything. Systematic observation across captive and wild groups has documented a repertoire of over a hundred distinct gestures, and — more importantly — has shown them meeting the standard criteria for *intentional* communication. The animal checks that a recipient is watching before signalling. It persists, or substitutes a different gesture, when the first one fails. It stops when it gets what it was after. Those three together are a demanding test that a great deal of animal signalling does not pass.

Much of the repertoire is shared across groups and across populations rather than being invented locally, which is what you would expect of something largely inherited rather than culturally constructed. So: intentional, flexible, sizeable, and species-typical. Not syntax, not symbols, not a lexicon — the meanings concern immediate goals.

The sign-language projects are a separate matter and are routinely summarised in whichever direction the writer prefers. What is not in dispute: apes in those projects acquired substantial vocabularies of manual signs and used them to obtain things. That is a real achievement and nobody has retracted it.

What did not survive scrutiny is the claim about grammar. The evidence for it came from teachers' contemporaneous notes, and the flaw sits precisely there: a human who wants an ape to sign will cue it — a glance, a hand position, a partial sign — and will not notice doing it, and notes recording the animal's signs cannot detect what happened in the second beforehand. Frame-by-frame analysis of videotape found that the great majority of one chimpanzee's signs reproduced a sign the teacher had just made or was still making. Apparent multi-sign utterances were largely repetition, showed no consistent word order, and did not lengthen with development as a signing child's do.

The best-known gorilla project reported vocabularies of many hundreds of signs, and is unresolvable in a specific way that should be stated rather than argued around: its primary records were never released for independent frame-by-frame analysis. That means the central claim cannot be checked in either direction, which is a different situation from having been tested and failed.

• Learned symbols — demonstrated. Apes acquire and use hundreds.
• Intentional communication — demonstrated, and in wild gorillas without any training.
• Reference to things not present — patchy, disputed, and design-dependent.
• Syntax — not demonstrated in any published ape corpus.
• Language in the human sense — a bundle of the above plus recursion and open-ended productivity, and no.

Keeping those separate is the whole discipline of this section. The frame-by-frame result is constantly cited as showing that apes are not clever, which it does not show and does not claim; and the vocabulary results are constantly presented as demonstrating language, which they do not. The honest position is that apes acquire vocabulary and have not been shown to acquire syntax, and that most of what the word "language" is doing lives in the difference.

On cognition more broadly, gorillas are the odd ones out among great apes in one respect: they generally fail mirror self-recognition tests that chimpanzees and orangutans pass. Whether that reflects a genuine cognitive difference or the fact that sustained eye contact is a threat in gorilla society — making the test itself hostile — has been argued for decades and is not settled.

The leaf-eating gorilla is a mountain gorilla. Combining feeding records and faecal analysis from six long-term western lowland sites gives a different animal: substantially frugivorous where fruit is available, with diets spanning more than a hundred plant species, falling back on leaves, stems and pith when the fruit runs out. Since western lowland gorillas are the overwhelming majority of all gorillas, the stereotype describes the minority.

On the meat question, which people ask often, the answer is specific: ants and termites are eaten in small quantities at most sites, and there is no established record of vertebrate predation. Insects are meat in one sense and not in the sense the question means.

Tool use by wild gorillas was first recorded only in 2005, at a forest clearing in the Republic of Congo where animals crossing swampy ground could be watched continuously. One female used a branch to test water depth ahead of her while wading; another laid a detached trunk across soft ground as a bridge and used it as a support while foraging. Both are postural uses — for balance and for information — rather than for getting at food.

That contrast with chimpanzees is the interesting part. Chimpanzees crack, fish and sponge routinely, because their food has to be extracted from things. A gorilla eating leaves and stems has nothing to extract. Captive gorillas had been using tools for far longer, so the rarity in the wild looks like a matter of ecological need rather than of capacity — and a capability that is never called on is indistinguishable, from the outside, from one that is absent.

Swimming is a good example of the distinction between does not and cannot. Gorillas avoid deep water, they are dense with low body fat, and zoos replaced their water moats after drownings — so as a description of gorilla behaviour, "cannot swim" is close to right. As a statement about physical possibility it goes further than the evidence: apes raised around water from infancy have been filmed swimming and diving deliberately, using a breaststroke-like leg action rather than the paddle nearly every other mammal uses, which the researchers relate to ape shoulders being built for climbing. Neither documented animal is a gorilla, and no gorilla has been tested.

A gorilla group is typically one mature silverback — sometimes several — with adult females and their offspring. What the silverback does is mediate conflict within the group, decide where it travels, and put himself between the group and a threat. Females associate with a particular male rather than with a territory, and can and do transfer between groups; a male who fails to protect loses his group rather than being deposed by it.

The alpha-and-hierarchy vocabulary is imported from a model of wolf packs that its own author spent decades retracting, and it fits gorillas no better than it fitted wolves. Aggression is directed at rival males and at perceived threats rather than being the medium of internal order, and the chest-beat display exists precisely so that contests between males can be resolved acoustically instead of physically.

That is not the same as gorilla society being gentle, and a benign summary would be dishonest. Infanticide by an incoming male is documented and is a real part of the system: a female whose infant is killed returns to fertility sooner. Naming it matters, because a page that corrects the "aggressive alpha" framing and then quietly omits this has traded one distortion for another.

The honest caveat on all of this is where it comes from. The overwhelming majority of gorilla behavioural observation is of habituated mountain gorillas — a population of around a thousand animals, in unusual habitat, watched daily for decades. Whether the same structure describes western lowland gorillas, whose groups are smaller and less cohesive and who are barely observed by comparison, is genuinely open.

Development is slow in the way that generally accompanies large brains. Infants nurse for around three years and remain with their mother for longer; females give birth roughly every four years; and gorillas live into their thirties and forties in the wild, longer in captivity. A silverback's back turns grey at around twelve to fifteen years, which is a marker of maturity rather than of old age.

All four gorilla subspecies are threatened, and one of them has a genuinely good story attached — which is rare enough in great ape conservation to be worth telling precisely. Mountain gorillas in the Virunga massif increased over decades of what the researchers themselves called extreme conservation: daily monitoring of every group, sustained anti-poaching patrols, and veterinary intervention for injured or sick individuals.

Comparing protected against unprotected groups identified the mechanism, which is the useful part. The difference was driven by reduced mortality — fewer deaths from snares and from disease — rather than by any increase in birth rate. That is a specific, purchasable outcome rather than a vague success.

The caveat is in the word extreme. This worked for one small population under conditions almost nowhere else can reproduce, at a cost per animal that would be impossible at scale. The far more numerous western lowland gorilla is not on this trajectory: Ebola outbreaks have killed very large numbers, commercial hunting continues, and the population is spread across an area no daily-monitoring programme could cover.

• How strong is a gorilla?

  Why it matters: It is the single most-asked question about the animal and has no measured answer. Everything in circulation is extrapolation from chimpanzee data or from 1920s trials.

  What would settle it: Fibre-level measurement of gorilla muscle, which would require tissue that is difficult and rarely appropriate to obtain.

• Why do gorillas generally fail mirror self-recognition when other apes pass?

  Why it matters: Either it marks a real cognitive difference among great apes, or the test is invalid for a species in which staring is a threat — and those imply opposite things.

  What would settle it: A self-recognition paradigm that does not require sustained eye contact with the reflection.

• Does the social structure described for mountain gorillas hold for western lowland gorillas?

  Why it matters: Nearly everything believed about gorilla society comes from about a thousand animals in unusual habitat, and they are a small minority of the genus.

  What would settle it: Long-term observation of unhabituated western lowland groups, which is extremely difficult in dense forest.

• Did the ape-language projects show any genuine syntactic sensitivity?

  Why it matters: The best-known gorilla case cannot be evaluated because its records were never released, so a central claim about ape cognition rests on data nobody outside the project has seen.

  What would settle it: Release of the primary records for independent frame-by-frame analysis.