Wild chimpanzees scaffold youngsters’ learning in a high-tech community

Across human history, the spiraling complexities of our technologies have been accompanied by a progressive elaboration in the schooling necessary to instill the skills that increasingly technological societies require. Among peoples who still subsist by foraging for wild foodstuffs using a toolkit that can be carried on one’s back, there is much to learn (1), but extensive formal schooling is unnecessary, as seems likely for our species’ long hunting and gathering past. By contrast, children in the societies that read this journal experience over a decade of schooling, and technical apprenticeships often last many years. Ape technologies, although much simpler than our own, also have been found, in recent years, to show much regional variation in complexity. Some populations display unimagined levels of sophistication in their manufacture and use of tools (2), significantly exceeding the complexity seen in other communities (3). Comparing 2 such relatively high-tech and low-tech communities, Musgrave et al. (4) report in PNAS that, by analogy with the human technology−education linkages sketched above, young chimpanzees’ social learning is more highly structured in the high-tech population, differing especially in the ways mothers offer costly support to the efforts of their offspring, which the authors class as an elementary form of teaching.

What counts as a relatively high-tech chimpanzee culture? Musgrave et al. (4) studied chimpanzees in the Goualougo Triangle, Republic of Congo, where previous studies have documented unusually sophisticated techniques to harvest termites (2). These contrast with the approaches documented at other locations across Africa, including Gombe, Tanzania, chosen as a relatively low-tech community for comparison (5). At Gombe, chimpanzees fish by inserting a single flexible probe into a termite mound, and utilize a variety of interchangeable materials for this, such as grass stems, bark, or twigs. At Goualougo, by contrast, chimpanzees harvest termites from the deep subterranean parts of nests using a “tool set” (6) necessary for success, and composed of more selectively chosen materials (2, 4). The process starts when a chimpanzee arrives at a fishing site carrying a stout stick of one particular species, or utilizes one discarded previously. This stick, which may be nearly a meter long, is pushed down into the ground, often helped by grasping the lower part with a foot, reminiscent of a person digging with a spade (Fig. 1A). The stick is intermittently retracted and sniffed, probably for cues of hitting a concentration of termites. It may be pushed nearly all of the way in, to create a deep tunnel down to the nest. This stick then is discarded, and the chimpanzee works with a second tool, a fresh stem (often one of several) that it has picked en route and held in the mouth. The stem is trimmed in length, leaves are stripped off, and the tip is then pulled through part-closed teeth several times to create a brush-like tip (Fig. 1B), which, as experiments show, is particularly effective for picking up termites in the nest. The tip is then moistened with saliva, and the fibers are compressed together to form a neat end, then efficiently threaded down the tunnel. Termites gathered are picked off the tool using the lips or swept off manually. Goualougo chimpanzees also fish in aboveground mounds, where a single fishing probe may suffice if a hole can be opened manually, but sometimes this also involves a tool set, with a perforating twig being used for the initial opening. Over 20 different forms of tool use have been documented at both Gombe and Goualougo, but the design complexity of the latter’s tool sets and associated techniques are more complex. At Gombe, a 5-y-old is typically a competent termite fisher, whereas subterranean termite fishing is achieved only close to adulthood at Goualougo (4).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Chimpanzee technology and the context of social learning in wild chimpanzees. (A) Creating a tunnel to fish for subterranean termites at Goualougo. Image courtesy of D. Morgan (Lester E. Fisher Center for the Study and Conservation of Apes, Chicago, IL) and J. Zampol (artist). (B) Creating a brush-tipped fishing tool by stripping through the teeth. Image courtesy of I. Nichols (photographer). (C) An infant watches termite fishing at Gombe. Image courtesy of K. Walker (North Carolina State University, Raleigh, NC). (D) A juvenile initiates fishing tool transfer at Goualougo. Image courtesy of D. Morgan (Lester E. Fisher Center for the Study and Conservation of Apes, Chicago, IL) and C. Sanz (Washington University in St. Louis, St. Louis, MO). For full descriptions, see text.

Musgrave et al. (4) suspected that an association might exist between these different levels of technological complexity and the nature of the social learning processes that support their acquisition, particularly concerning mothers’ actions facilitating the development of their offspring’s termite fishing skills. A variety of quantitative measures extracted from videos recorded at the 2 study sites offered supporting evidence for this hypothesis (4). Youngsters at both study sites avidly watch expert termite fishers in action (Fig. 1C), typically their mothers, and such close peering offers ample opportunity for observational learning, increasingly recognized to pervade the lives of apes (7, 8). However, harvesting and manufacturing good fishing probes appears more challenging than fishing itself, so, if youngsters can scrounge good fishing tools, they can begin to practice fishing. It is the maternal role in this that the authors show to differ greatly between the communities. Transfers of manufactured tools from mother to offspring occur at Goualougo at over 3 times the rate seen at Gombe, and the probability of transfer after the youngster requests a tool (usually by reaching for it and whimpering: Fig. 1D) can be 5 to 8 times higher at Goualougo. Moreover, the authors recorded a category of active transfer following such a request that occurs only at Goualougo, where it is the most common type of response. In these instances, following an offspring’s request, a mother moves so as to actively offer her tool—the tool is not simply filched from her. In some of these cases, the mother neatly splits the probe she was using down the middle, creating one viable tool for her offspring and another she can continue to use herself. By contrast, a refusal to transfer a tool following a request was 3 times more common at Gombe.

Should such behavior in Goualougo mothers count as “teaching,” as the authors’ title proclaims (4)? When we humans talk of teaching, we generally think of it as defined by intent—having the intended purpose of educating another individual in some way. That is not something that Musgrave et al. can demonstrate, nor do they attempt to do so. Instead, they appeal to a definition of teaching proposed earlier by Caro and Hauser (9), that has achieved wide currency in the field of animal behavior. These authors offered a conception of teaching that identified it on the basis of its beneficial consequences, reflecting the adaptive function it appears to have evolved to serve, likely shaped by natural selection. This approach has appealed to ethologists, who are inclined to focus on the biological function of an activity, aside from questions about its intent. Caro and Hauser accordingly suggested that an activity with a teaching function could be identified by a cluster of criteria, notably, the putative teacher modifying its behavior in the presence of the pupil in a way that provides no immediate benefit but serves to promote learning by the pupil. Not all in the field are comfortable with setting aside the issue of intent in this way (10), especially when any comprehensive analysis of the evolution of teaching is at stake, which must include humans. Nevertheless, teaching defined in functional terms has been identified in many different animal species (11, 12), as

Research on cultural processes such as those studied by Musgrave et al. may be turned to use in efforts to sustain the cultural and behavioral diversity of our highly endangered, closest animal relatives in the forests of Africa.

when adult meerkats provide youngsters with scorpions to practice on, after initially removing stings, but later leave the prey intact as competence in killing grows (13). As Musgrave et al. argue in their present paper and in an earlier analysis (14), a Goualougo mother’s actions appear to meet such functional criteria insofar as they enhance opportunities for skill development in her offspring, while also imposing an apparent cost to herself, as she has to obtain and shape a new tool of her own (al though a mother who ingeniously splits her tool neatly minimizes such costs!).

I suspect that, nevertheless, some researchers in the field will feel that even the notion of functional teaching may here be stretched too far. The approach that Caro and Hauser (9) advocated perhaps was underlain by an implicit assumption that candidate actions would be active initiatives on the part of putative teachers, consistent with teaching’s manifestations in humans and with the examples studied most by Caro and Hauser, such as felines bringing dead, then later live, prey to their offspring, or primate mothers punishing infants emitting alarm calls inappropriate to the context. By contrast, it is youngsters who initiate transfers at Goualougo, to which mothers are then especially responsive. This might be thought more akin to an enhanced tolerance of scrounging, or curtailing of intense offspring pestering. However, whether or not talk of teaching is justified [“scaffolding” might be a more justifiable term, employed in the authors’ text but not their title, and used in developmental psychology (15) to label functions like those described in the chimpanzees] is arguably less important than the notable contrasts between Gombe and Goualougo in maternal behavior. These are consistent with, and indeed provide circumstantial evidence for, some coevolutionary linkage between the sophistication of technology in a nonhuman society and the complexity of the social learning processes that support its cultural inheritance.

This raises the question of what may cause any such correlation. One obvious candidate is a genetic difference between the mothers at the 2 sites, that may have been selected for by the differing technological demands at the 2 study sites. Maternal scaffolding instead might have arisen through cultural evolution and be transmitted across generations by processes of social learning, perhaps a candidate nonhuman example of what Heyes (16) has called a “cognitive gadget”—a cognitive phenomenon shaped by culture rather than genetics. Alternatively, gene−culture coevolution may have occurred, a phenomenon for which there is increasing suggestive evidence across nonhuman species, as for humans (17). Perhaps such speculations are absent from Musgrave et al.’s (4) discussions because their empirical resolution appears currently beyond our grasp.

A skeptic might, in any case, object that the contrast between the study sites involves just 2 chimpanzee communities; in other words, a graph of technology sophistication against scaffolding complexity, although consistent with the existence of a correlation, is essentially based on just n = 2 data points. Clearly, more cases are needed before secure confidence can be placed in the interesting indications of coevolution. Tragically, recent evidence suggests that human environmental degradation is impacting extremely negatively on the very diversity of chimpanzee behavior on which such investigations are based (18), leading to calls to take account, in conservation efforts, of cultural phenomena and the diversity they engender (18, 19). In this way, research on cultural processes such as those studied by Musgrave et al. (4) may be turned to use in efforts to sustain the cultural and behavioral diversity of our highly endangered, closest animal relatives in the forests of Africa.