Following the Gardners' discovery that an ape named Washoe could learn to produce and combine a number of hand movements similar to those used by deaf human beings, a variety of `ape-language projects' sprang up. Some projects used different symbol systems, others used different training techniques, and others used different species of apes. While debate still rages regarding the appropriate way to interpret the symbolic productions of apes, three species of great apes (gorilla, orangutan, and chimpanzee) have now been credited with this capacity while no lesser apes or monkeys have been reported, at present, to have acquired such communicative skills. Among all of the claims made for the various animal species, the philosophers have entered the fray attempting to define the essence of what it is about language that makes it `human'. This paper will compare and contrast the above positions to arrive at behavioural definitions of symbolic usage that can be applied across species. It will then present new data on a fourth ape species Pan paniscus which is proving to be the first non-human species to acquire symbolic skills in a spontaneous manner.