Introduction

The primary aim of research into artificial intelligence is to replicate biological capabilities either through attempting to directly copy biological mechanisms or by engineering solutions from non-biological principles. Either way, the first hurdle is to replicate the behaviour of animals. The biological phenomenon of mimicry includes a range of behaviours that are of particular interest because they represent a signal–receiver relationship in which the interests of both signallers and receivers are much more clearly defined than in many communication systems. Unlike examples of communication such as mate attraction where it is far from clear what the best evolutionary interests of signaller and receiver are, in mimicry it is generally clear that the mimic aims to deceive the receiver and the receiver aims to avoid being deceived. A second interesting feature of mimicry is that it is possible for human observers to estimate how accurate particular examples of mimicry are through their own (albeit subjective) observations. It is striking that examples of mimicry vary substantially in their accuracy, providing information about the selection pressures acting on the evolution of the trait.

Mimicry is widespread in nature and the term encompasses diverse behaviours, morphologies and/or capacities involving three parties (of up to three species): the imitator (mimic), the imitated (model) and the recipient (signal receiver). These terms may refer to one organism resembling another in order to fool a third (as in the familiar butterfly mimics), in other situations they may refer to mechanisms that aid an individual's reproductive success (as in female impersonation by sneaker males), in others again it may refer to complex learned and acquired abilities used to pass skills from one generation to the next within a species (as in primate learning).