Journal of Experimental Psychology: Gei1986, Vol. 115, No. 3,211-235Copyright 1986 by the American Psychological Association. Inc.0096.3445/86/$8o.75Spontaneous Symbol Acquisition and Communicative UseBy Pygmy Chimpanzees (Pan paniscus)Sue Savage-RumbaughDepartment of Biology, Georgia State UniversityYerkes Regional Primate Research Center, Emory UniversityKelly McDonald, Rose A. Sevcik, and William D. HopkinsLanguage Research Center, Georgia State UniversityElizabeth RubertYerkes Regional Primate Research Center, Emory UniversityTwo pygmy champanzees (Pan paniscus) have spontaneously begun to use symbols to communicatewith people. In contrast to common chimpanzees (Pan troglodytes) using the same communicativesystem, the pygmy chimpanzees did not need explicit training in order to form referential symbol-object associations. Instead, they acquired symbols by observing others use these symbols in dailycommunications with them. In addition, the pygmy chimpanzees have begun to comprehend spokenEnglish words and can readily identify lexigrams upon hearing the spoken words. By contrast, com-mon chimpanzees who received similar exposure to spoken English are unable to do so. The olderpygmy chimpanzee has begun to form requests of the form agent-verb-recipient in which he isneither the agent nor the recipient. By contrast, similarly aged common chimpanzees limited theirrequests to simple verbs, in wihch the agent was always presumed to be the addressee and the chim-panzee itself was always the recipient, thus they had no need to indicate a specific agent or recipient.These results suggest that these pygmy chimpanzees exhibit symbolic and auditory perceptual skillsthat are distinctly different from those of common chimpanzees.What's in a Name?The language acquisition capacity of apes has been the focusof a number of research projects (Asano, Kojima, Matsuzawa,Kubota, & Murofushi, 1982;Fouts, 1972; Gardner & Gardner,1971; Hayes & Hayes, 1951; Kellogg & Kellogg, 1933; Miles,1983; F.Patterson, 1978;Premack, 1971;Rumbaugh, 1977;Sa-vage-Rumbaugh, 1984b, 1986; Terrace, 1979). Despite thedifferences among various projects regarding the communica-tive systems (American Sign Language [ASL], geometric sym-This research was supported by National Institutes of Health GrantsNICHD-MOI6 and RR-00165 to the Yerkes Regional Primate Re-search Center, Emory University.The authors wish to thank Duane Rumbaugh for his constant help,advice, guidance, and colleagueship throughout the study; without himthe work never would have been possible. His comments on variousdrafts of the manuscript also made it a much better article.We also wish to thank Mary Ann Romski for being such a good friendto Kanzi, and for conducting the blind field test, in spite of the feet thatshe was not fond of woods. Her keen attention to Kanzi's communica-tions are responsible for the success of the test.Jeannine Murphy and Phillip Shaw worked daily with Kanzi andMulika throughout this entire period and helped them learn the manythings described in this paper. They were, and are, essential to the wellbeing and development of Kanzi and Mulika.Correspondence concerning this article should be addressed to SueSavage-Rumbaugh, Yerkes Regional Primate Research Center, EmoryUniversity, Atlanta, Georgia 30322.bols, or plastic tokens) used, it is generally agreed that apes canlearn to produce symbols in order to bring about the occur-rence of desirable events (Jolly, 1985; Savage-Rumbaugh,1984b; Terrace, 1985).It has also been shown that when a chimpanzee is systemati-cally taught the communicative skills of requesting, labeling,and comprehending, indicative referential symbol usage canthen appear without further training (Savage-Rumbaugh,1984b, 1986; Savage-Rumbaugh, Pate, Lawson, Smith, & Ro-senbaum, 1983). With the appearance of indicative usage, apesgo beyond conditioned request-oriented utterances and becomeable to make statements about behaviors they are going to en-gage in, in advance of actually emitting those behaviors. Theability to use symbols to orient the attention of others to partic-ular aspects of the environment, other than the fulfillment ofimmediate needs, also appears to accompany the onset of indic-ative communication (Savage-Rumbaugh, 1984b, 1986; Sav-age-Rumbaugh et al., 1983).Nonetheless, it is still the case that the manner in which apeslearn language skills differs from the way in which the humanchild learns language. That is, apes have generally required spe-cial training sessions during which they are reinforced for pro-ducing the correct symbol in the presence of a particular object.During such training sessions, they are repeatedly shown an ob-ject, or engaged in an activity, then required to name that objector that activity (Fouts, 1972; Gardner & Gardner, 1971; Pre-mack, 1976; Rumbaugh, 1977).One ape, Loulis, (an infant adopted by Washoe) has been ob-served to produce a number of hand gestures without specific211212SAVAGE-RUMBAUGH, MCDONALD, SEVCIK, HOPKINS, RUBERThuman training (Fouts, Hirsch, & Fouts, 1982). However,Loulis has not been tested to determine whether he in fact un-derstands the referents of any of these gestures. Because it is welldocumented that both captive and wild chimpanzee popula-tions learn various nonreferential gestural signals (McGrew &Turin, 1978), it may be that Loulis' gestural capacities do notfunction any differently from those of other captive apes whouse simple nonreferential gestures (Savage-Rumbaugh, 1984a,1986; Savage-Rumbaugh et al., 1983; van Lawick-Goodall,1968). In any case, until Loulis's abilities are appropriately as-sessed, we cannot conclude that he has learned referential signs.It is possible for apes, and for language-disabled children, touse gestures appropriately, without being able to comprehendthe gestures at the referential level. In such cases it can be ar-gued that the individuals in question have not made the transi-tion from associative naming to representational naming (Rice,1980; Savage-Rumbaugh et al., 1983). Similarly, the earliestgestural and vocal symbol usage in normal children is primarilyassociative. It is only in a gradual manner that words becomerepresentational and context independent (Lock, 1980).It is this transition from associative to representational sym-bol usage that is the critical issue for the proper understandingof language use in both apes and children. As Lock (1980)put it:The child's early learning of words presents two main