Animal communication explained: how signals actually work, and the examples that show why

Animal communication is the transfer of information from one animal, or a group of them, to one or more others, in a way that affects the current or future behaviour of the receivers. The definition is deliberately behavioural. It does not ask what the sender intended or what the receiver understood, because neither is observable. It asks whether the receiver did something different.

That framing does real work, because it separates communication from information leakage. Information can move without anyone signalling. A predator leaves scent, its prey detects the scent and flees: information has transferred, via what are called kairomones, but the predator gained nothing and would prefer it had not happened. Nobody signalled.

What animal communication is, explained

The technical term for information that does change a receiver’s behaviour is a signal, and signalling theory sets a hard condition on which signals persist. For a signal to be maintained in a population, both the sender and the receiver should usually get some benefit from the exchange. If only the sender benefits, receivers evolve to ignore it. If only the receiver benefits, senders stop paying for it.

This is why so much of the field is preoccupied with honesty. The classic case, explained in every textbook, is the peacock’s tail. One account, the good genes hypothesis, treats an elaborate display as an honest advertisement: the bird really is a better mate. The other, the handicap hypothesis, turns the cost into the mechanism. A tail takes energy to maintain and makes its owner easier for a predator to see. Precisely because it is expensive, it cannot be produced by an animal in poor condition, which makes it a reliable indicator whether or not the sender is trying to be honest.

How do animals communicate?

Through whatever channel physics makes available. Ask how do animals communicate and the answer divides by medium rather than by species, because unrelated animals converge on the same channel when the environment rewards it.

Chemical. Many social insects release pheromones to lead others to food. Ants lay a pheromone trail on the ground, and other ants follow it. The signal is external to the animal, persists without further effort, and decays on its own, which makes it a good solution for a colony that needs a route to stay usable for a while and then stop being usable.

Acoustic. Vocal signalling covers advertisement against rivals and towards mates in frogs, hammer-headed bats, red deer, humpback whales, elephant seals and songbirds. It also carries information that is more specific than mood. Campbell’s monkeys have alarm calls; gibbons have territorial calls; greater spear-nosed bats use frequency to tell groups apart.

Tactile. Some ants recruit by physical contact: patrollers leave the nest to check for danger and return to recruit foragers by touching them. The honeybee waggle dance belongs here too. And in colonial animals such as corals, individuals are linked closely enough that an entire colony can react to an alarm movement made by only a few members.

Visual, including light. Some signals are colour changes tied to physiological cycles: when a female olive baboon begins to ovulate, her anogenital area swells and turns bright pink, which signals readiness to mate. Where there is no ambient light to reflect, animals make their own. Bioluminescent communication is common in the oceans, particularly at depth, and Humboldt squid can signal visually in water that is effectively dark.

Channels and examples that show the limits

The most-cited example in the whole field is worth stating precisely, because it is usually overstated. Vervet monkeys give a distinct alarm call for each of their four different predators, and other monkeys’ reactions vary appropriately with the call. A call that signals a python sends them into the trees. This is not general-purpose language. It is a small set of specific referents, which is a much stronger claim than “they sound agitated” and a much weaker one than “they have words”.

Two further examples that complicate any simple account:

• Ravens signal to older ravens when they encounter new or untested food, which puts the benefit on the receiver’s side and raises the question of what the sender gets.
• Rhesus macaques produce food calls to inform other monkeys of a food source, and do so to avoid punishment. The signal is honest, and the reason it is honest is enforcement.

Autocommunication: when the sender is the receiver

There is a category where the model breaks in an interesting way. In autocommunication the sender and the receiver are the same individual: the animal emits a signal, the environment modifies it, and the same animal receives the modified version back. Echolocation is the familiar case. The information is not in the signal that was sent. It is in the difference between what was sent and what came back.

Perception more broadly can carry signals nobody sent. Some snakes detect infrared well enough to localise warm-blooded animals such as cattle and horses to within roughly 10 to 15 centimetres, and that sensitivity may be used to find the regions of highest blood flow on a target. That is information transfer with no signaller at all, which is exactly the case the behavioural definition of animal communication is built to exclude.

Communication that runs in the other direction

The examples above are animals signalling to animals. The reverse case, animals producing signals deliberately aimed at humans, is a harder experimental problem, because the researcher has to rule out the animal responding to unintentional cues from the researcher. Work of that kind has been done with horses in Norway, and it is the subject of a separate piece on horses trained to indicate a preference.