A third call seems to serve to keep the members of a group close enough together as they move through the woods in search of food. The most important property of this system of communication is its lack of flexibility. Whatever the exact number of calls is, it is finite and small. Whilst any one of the calls can be varied in loudness or in the number of repetitions, no matter what situation a gibbon may encounter, its vocal reaction is constrained to be one of this small finite number: a gibbon does not react to a new situation by producing a new call or putting together two or more of the calls already available.

It has been suggested that language grew out of a primate call system, like the ones used by apes today and the one I have described that is used by gibbons. It is assumed that humans started out with a simple set of cries in which each one meant something different, such as, "Danger!" or "Follow me!". These cries gradually became more elaborate and eventually evolved into language. A possible intermediate stage is seen in the cries of the vervet monkey. This monkey has several alarm calls which distinguish between different types of danger. According to some, it is a very short step from an alarm call warning of a poisonous snake to using the chutter as a 'word' signifying a poisonous snake. However, another interpretation of these signals is possible; a distinguishing between the intensity of different types of danger. However an experiment in which a concealed loudspeaker plated recordings of the alarm calls showed that when each sound was heard, a specific response was seen.

For example, when they heard the chutter, the vervets stood on their hind legs and looked around for a snake. This implies that the monkeys clearly have a special signal for each type of enemy; each signal has semantic properties.

It is perhaps unfair to concentrate on primates. Compared with these, bees and dolphins have extremely sophisticated communication systems. Whilst the majority of bee communication is achieved through dance rather a system of sound signals, I feel that there is some evidence in this example that casts some degree of doubt on the supposed human-specificity of language. When a worker bee finds a source of nectar, it returns to the hive and performs a dance. It has been shown that certain features of the dance transmit information about the location of the source of nectar: one feature specifies the direction of the source from the hive in relation to the position of the sun, another specifies its distance.