(Man 24:383-398; 1989)

[This is a scanned version that may include scanning errors]



University of Colorado at Colorado Springs



University of Stirling


When in human evolution did our ancestors cease behaving like apes? In this article we address this question by interpreting the earliest known archaeological evidence, the Oldowan, in light of what primatologists know about modem apes, especially the chimpanzee (Pan troglodytes). Our analyses consider aspects of Oldowan tools and tool-making and those aspects of Oldowan subsistence that can be reconstructed from artifacts. We conclude that all the behavior that can be inferred from Oldowan tools and sites falls within the range of the ape adaptive grade. There is nothing exclusively human-like about this oldest known archaeological evidence. However, the Oldowan did include two specific behavioral patterns that, while still within the ape adaptive grade, are almost unknown for modern apes and which point in the direction of adaptations found later in hominid evolution. These are carrying tools or food for thousands of meters and competing with large carnivores for animal prey.




The question of the origins of culture is a central one to anthropology. While anthropologists acknowledge an evolutionary continuum with apes, we also recognize a huge difference between the cultural behavior of modern apes and the cultural acts of modem humans. When in human evolution did our ancestors cease behaving like apes? Humans and apes have been phylogenetically separate for several million years, but this does not mean that our ancestors had human-like behavior for the same length of time. The anatomy of fossil hominids cannot fully answer this question because it tells us little about the kinds of behavior at issue. We must turn to the archaeological record.

The Oldowan Industrial Complex is the oldest known flaked stone technology. In east Africa tools from this complex date from as early as 2.5 million years ago (Harris 1983) and by 2 million years ago they occur in archaeological sites associated with broken up animal bones (Leakey 1971: Isaac 1976). The archaeological record for the Oldowan is meager, however, when judged by the standards of more recent prehistory. From a few thousand tools and animal bones, and their geological context, archaeologists must try to reconstruct the behavior of a tool-maker generally assumed to be a precursor of modem humans. Because there are so few pieces in this archaeological puzzle, we must use other sources to complete the picture of Oldowan life. This is largely a matter of selecting appropriate analogies.

The traditional view of the Oldowan is one of a proto-human culture, a first step on the evolutionary trail leading to modem humans. Archaeologists have often used knowledge of modem hunters and gatherers as a source for understanding the archaeological patterns of the Oldowan (e.g. Isaac 1978). However, when one looks ‘backward’ to the Oldowan from the perspective of modern humans, or even from the perspective of recent Stone Age humans, one risks filling in the missing pieces with inappropriate analogies. The result may be an over-optimistic interpretation. For example, Parker and Gibson (1979) have attributed several behavioral patterns to Oldowan hominids such as aimed throwing, dividing carcasses into equal portions, and sharing. There is no tangible evidence for any of these in the archaeological record, but they are well known for recent hunters and gatherers. In a similar vein, Tobias (1983) has argued for a ‘holistic picture of the lifestyle’ that included language along with the behavior cited above. Neither the fossil record for hominids nor the archaeological data give direct evidence for language. Tobias presented this picture as a reasonable interpretation. And it is, if one assumes that the Oldowan was proto-human.

In the following analyses we take a different perspective. We look for what is ‘ape’ about the Oldowan. Based on knowledge of the modern Pongidae (the great apes), especially the chimpanzee, Pan troglodytes, we examine the archaeological patterns presented by the Oldowan tools and sites (cf. Tanner 1987).

We try to match these patterns with specific kinds of behavior known for modern apes and to assess if such behavior is congruent with the foraging, technological and mental competencies of apes in general, something we term an ape adaptive grade. Reference to an ape adaptive grade is necessary because many specific differences, especially in terms of subsistence, are to be expected between hominoids occupying different niches. However, the basic types of solution to adaptive problems are likely to be much the same, with the differences being merely variations on a theme.

At the outset we must mention a logical difficulty faced by all archaeologists: the problem of minimum necessary competence. In prehistory, one can assess only the minimal abilities needed to produce the archaeological patterns uncovered. This is especially maddening for the Oldowan, because, as we seek to show in this article, the minimum competence needed to produce Oldowan tools and sites is that of an ape adaptive grade. One solution to this problem is to assume the existence of behavior that would have required a more human-like competence, but which is not evident in the archaeological record. “We see no reason to assume the existence of such behavioral patterns, as the point of scientific archaeology is to use evidence, not make assumptions. Thus, for the purposes of this exercise, we accept minimal competence as being a reliable description of Oldowan behavior.

The article is organized into two main sections. The first addresses Oldowan tools and their counterparts in ape behavior. The second similarly treats the patterns of tool-use. Then follows a brief summarizing conclusion and a more speculative discussion.




The most direct way to compare Oldowan and chimpanzee technology is to compare the tools themselves (see table 1 and fig. 1). This approach has the advantage of dealing with more or less tangible phenomena. Here we will emphasize the cognitive prerequisites of the various tools. In assessments of the Oldowan the general presumption has been that it was a comparatively complex technology based on complex cognition (see, e.g. Gowlett 1984 or Toth 1985). After first considering the raw materials, we will examine the spatial concepts used in tool-making, the natural groupings of the tools and, finally, the manufacturing procedures.

All known Oldowan tools were made of stone; most chimpanzee tools are not. This distinction has maintained the status of a phylogenetic divide despite much documentation of ape tool-use to the contrary. We question whether or not this simple difference in raw material deserves such significance. Chimpanzees use stones as weapons in antagonistic encounters and in processing plant foods. True, chimpanzees do not flake stone, though they occasionally produce unintentional flakes when they crack nuts on stone anvils. These result from mis-hits when a hammer-stone clips the edge of the anvil (Hannah, unpublished data); apes have yet to be seen to use such flakes. There is reason to suppose, however, that the use of intentionally flaked stone is within their range of abilities. A male orangutan successfully completed an experimental task in which he had to strike a flake from a core using a hammer-stone and then use a flake to cut through a cord securing a box (Wright 1972). More recently, Kitahara-Frisch et al. (1987) reported that chimpanzees shown how to use hammerstones to smash long-bones then used the resulting bone fragments to puncture a skin covering a bottle of sweetened drink. Thus, the lack of functional stone flaking in wild chimpanzees does not represent some behavioral or cognitive deficit, but rather a difference in the raw material used for most tools and, perhaps more importantly, a difference in the tasks performed by the tools. We must look elsewhere for evidence of greater complexity in the Oldowan.


TABLE 1. Oldowan and chimpanzee tools compared.


Chimpanzee 1

Oldowan 2

1.   Known subsistence tools

termite probes

ant probes

leaf sponges

wooden hammers

stone hammers

stone anvils

tree-root anvils

1.   Known tools

stone hammers

stone anvils

flaked cores

unmodified stone flakes

modified scone flakes


flaked bone


2.   Spatial concepts

Primitive spatial concepts of proximity, boundary, and order used to make tools of vegetation.

2.   Spatial concepts

Primitive spatial concepts of proximity, boundary, and order used to flake stone.


3.   Natural groupings of tools

Some standardization results from selection of raw material of appropriate size and strength for the task. Bitten to appropriate length.


An ‘ad hoc’ technology.


3.   Natural groupings of tools

Some standardization results from selection of lava for large tools and quartz and quartzite for small tools An ‘ad hoc’ technology.

4.   Manufacturing procedure

selectivity of raw material modification by stripping, biting, chewing


trial-and-error sufficient

4.   Manufacturing procedure

selectivity of raw material modification by flaking



trial-and-error sufficient


1.   (Boesch & Boesch 1983, 1984; Goodall 1968; Hannah & McGrew 1987; Hasegawa et. al. 1983; McGrew 1974; McGrew & Collins 1985, Sabater Pi 1974; Sugiyama 1995, Sugiyama & Koman 1979; Teleki 1973).

2.   (Bunn 1981, Harris 1983; Isaac 1976; 1981; 1984; Leakey 1971; Ohel 1984; Potts 1984, Toth 1985; Wynn 1978. 1981, 1989).


One of the more direct ways to assess the cognitive ability employed in tool-use is through examination of spatial concepts. The advantage of this approach is that we can focus on the tools as products, particularly the geometric arrangement of elements on the tools, and do not have to see the sequence of manufacture or use. The spatial concepts required for Oldowan tools are primitive. The maker need not have paid any attention to the overall shape of the tool; instead, his focus appears to have been exclusively on the configuration of the edges (Isaac 1981; Toth 1985). As a consequence one need not argue for the use of such relatively sophisticated ideas as perspective or symmetry. Examination of the arrangement of trimming on the edges suggests a fairly consistent set of three spatial notions, all of which fall under the purview of topological space. These are proximity, boundary and order. The simplest of these notions is proximity or ‘nearby-ness’. By repeatedly bashing a cobble in more or less the same place knappers could create a number of sharp projections and intersecting edges. Placing, blows in proximity is a simple concept but it does require some coordination of motor patterns.


Figure 1.     Sample of Oldowan flaked stone tools.  Very simple concepts of space would suffice for positioning all of the modifications.  These modifications altered the shapes of edges but apparently without any attempt to achieve an overall shape.


A more complex notion is that of a boundary in space. A boundary is a feature that divides a spatial field into two realms or areas. Bifacial flaking requires respect of a boundary. By trimming on alternating sides of a boundary on a cobble the Oldowan tool-maker could produce a single sinuous edge; the boundary in a sense divided the spatial field of the cobble into two realms.

The most sophisticated spatial notion used on Oldowan tools is that of order. When making unifacial scrapers the tool-maker had to coordinate proximity with a notion of constant direction of movement: one trimming blow after another along the edge of a flake or core. The result is a sturdy, relatively even edge.

Very few Oldowan artifacts have discontinuous sections of trimming. It appears that the initial trimming had a tyrannical control on later trimming. The first blow seems to have anchored the rest; concepts of proximity, boundary and order then extended the trimming from this starting point. For example, when making a bifacial chopper the first blow determined the location of the boundary that guided subsequent trimming. All in all, the patterns of trimming on Oldowan tools are very simple and indicate that hominids used simple spatial concepts to coordinate their stone knapping (Wynn 1981; 1989).

Unfortunately, a similar assessment cannot be applied to most chimpanzee tools. Many of their characteristics are largely determined by the raw material: the position of the leaves to be stripped from a twig is dictated by the plant, not the ape. However, all the conceptual or perceptual abilities needed for Oldowan tools appear in tools used by apes. Rigorous experimental studies have shown that when given a blank sheet of paper upon which to draw, chimpanzee ‘artists’ do not mark randomly. When given a sheet upon which a single geometric shape like a square is already present, they concentrate their marks on and around it, thereby showing the property of proximity (Smith 1973; Boysen et al. 1987). The property of boundary is exemplified in an unusual kind of termite-fishing tool made only by the chimpanzees of Bilenge in Tanzania (McGrew & Collins 1985). The raw material is a sedge (Cyperus pseudoleplocladus), the stem of which is triangular in cross-section. For use as a fishing-probe, one of the three ridges is removed by careful longitudinal stripping, leaving a thinner more flexible tool which is trapezoidal in cross-section. Order is exemplified in the building of sleeping platforms (‘nests’), something which all great apes in nature do every evening. For chimpanzees, the first major branch bent over determines the foundation, which is followed by the interweaving of side-branches and finally lining of the nest with detached leafy twigs (Goodall 1968). Unlike the stereotyped nest-building of songbirds, this construction by apes requires appropriate experience in ontogeny if it is to be manifest (Bernstein 1962).

In sum, all the spatial concepts needed for Oldowan tools can be found in the minds of apes. Indeed, the spatial competence described above is probably true of all great apes and as such does not make Oldowan tool-makers unique.

Another approach to the cognitive underpinnings of tools is through scrutiny of the natural groupings into which they fall. Do the tools fall into relatively discrete categories or kinds, or do they show a more continuous variation in features like size and shape? Such a search for natural ‘types’ is a perennial problem of archaeological methodology precisely because there is no way to know that the categories which we recognize in archaeological assemblages correspond to the mental categories of the artisans. As a result, many archaeologists avoid the question altogether by using problem-specific analytical types, for which no argument for correspondence is needed. Nevertheless, from the perspective of the evolution of cognition, the nature of natural groupings is important. How boundaries between groups are demarcated, and perhaps more significantly, how the resulting categories relate to one another tells us some interesting things about how the artisans organized their world. Such matters are at the core of behavior and hence are worthy of investigation.

Leakey (1971) has argued that Oldowan tools fall into about two dozen types, including such types as side-choppers and end-choppers. While Leakey’s familiarity with the artifacts is unmatched, it is unclear whether or not these types and sub-types represent categories that existed in the minds of their makers. If one lays out the artifacts on a table there is almost continuous variation in size and shape (Wynn 1978). The one clear discontinuity is between larger artifacts on lava cores and smaller artifacts on quartz flakes. Isaac (1984) and Toth (1985) saw this continuous variation as resulting from production of edges necessary to complete a task immediately to hand. In effect, the individual had a specific task to perform. chose a blank of the right size (or perhaps made one in the case of flakes) and chipped an appropriate edge. After finishing the task the tools were abandoned. The result of such an ‘ad hoc’ technology would be a relatively continuous range of size and edge configurations corresponding to the range of specific tasks at hand. This interpretation better matches the variation seen in artifacts than does Leakey’s claims for types and sub-types. From a more cognitive perspective, it seems that Olduvai’s tool-makers need not have employed any standardized set of designs in the form of ‘mental templates’, or need not have used any concepts of types or sub-types.

Chimpanzees’ tools present a pattern of variation that is directly comparable. The size and shape of chimpanzee tools seem to be tied to the task at hand. In making probes to ‘fish’ for termites, chimpanzees at Gombe in Tanzania select aptly sized stems or grass blades and often bite them to optimize their dimensions or to rejuvenate the stiffness of the end (Goodall 1968). When faced with a different task, of dipping for driver ants, chimpanzees bite the probes to a different length (McGrew 1984). In the Okorobiko mountains of Equatorial Guinea chimpanzees break open the termites’ mounds rather than probe them, but again the tools are fairly uniform in features (Sabater Pi 1974). The uniformity of the artifact is governed by the uniformity of the task; it is still ad hoc technology. There is no reason to hypothesize a pre-existing image of an ‘Ideal’ probe. One makes what will work at the time. There is also no evidence of types and sub-types in chimpanzees’ tools. In sum, the chimpanzees’ apparent lack of mental templates, formal classification and pre-existing types matches the kinds of tool-groupings seen in the Oldowan.

While we cannot watch Oldowan tools being made, we can identify some of the steps in the procedure. Two are of special interest because they appear on the surface to be sophisticated: selectivity of raw material and the use of tools to make tools. In the Oldowan assemblages from Olduvai Gorge there is a marked tendency of the smaller tools to be made of quartz and quartzite and for the larger tools to be made of lava (Leakey 1971; Ohel 1984). The quartz and quartzite had to be carried for a distance of at least two kilometers. Thus, it seems that for reasons unknown the tool-users selected certain kinds of raw material for certain kinds of tools. This in turn implies a certain amount of foresight and suggests that tool-use was not entirely spur-of-the-moment. The knappers also needed to use stone hammers to make the flaked stone tools; in other words they used tools to make other tools. This striking point has achieved some notoriety in discussions of the evolution of intelligence (e.g. Parker & Gibson 1979) and has become a kind of threshold marker dividing ape from human technology.

From an ape’s perspective neither of these is remarkable. Chimpanzees in the Tai Forest in Ivory Coast regularly use harder hammers of granite and quartzite on Panda nuts than on Coula nuts, which can be opened by softer hammers of laterite (Boesch & Boesch 1983). Moreover, they regularly carry such hammers to nut-cracking stations, sometimes for distances of over 500 meters. Such selectivity is directly comparable to that at Oldovai.

Apes in nature have not yet been seen to use tools to make tools. (Sugiyama, (1985), has found chimpanzees in Cameroon using a ‘brush-stick’ to collect termites. It looks as if the frayed end of the stick has been beaten with a hammer, but the making of the tool remains to be directly observed). But is this such a significant failing, Two considerations argue against ‘tools to make tools’ as the great Rubicon of technological evolution. First, as noted above, it is possible to teach apes to use stone hammers to make tools with cutting edges; the procedure is well within ape capacities. Second, apes have efficient canines and incisors and for most of the tools that apes make these teeth are enough. However, one clearly cannot bite stone, so teeth alone would not suffice to make the tools used by Oldowan tool-makers. We cannot fault apes for not employing unnecessary techniques in making the tools needed in their subsistence.

Perhaps more enlightening than point-by-point comparisons between Oldowan and ape tool-making is a general comparison of the procedures used by both. For the Oldowan this means describing the least complicated procedure that would suffice for accomplishing the task (recalling that we must use minimum competence as a benchmark). Even the most complex Oldowan tool could be produced through trial-and-error. The tool-maker need only have had an image of the intended result such as a chopper and a repertoire of motor patterns for achieving it. She made a few modifications with trimming blows, checked the result to see if it would work for the task at hand, made a few more, checked again, and so on until she had produced an acceptable tool. Trial-and-error is not a haphazard, serendipitous kind of operation as the term may imply, but a procedure that employs well-defined intentions and techniques, What trial-and-error lacks is the use of contingencies for dealing with unanticipated problems. Moreover, trial-and-error procedures can control for only one variable at a time; that is, it cannot take account of the effect that an action may have on several variables at once or of the inter-relations between several variables (Piaget 1972; Wynn 1981). There is nothing about Oldowan tools that would require simultaneous consideration of several variables and nothing that would require contingency plans.

Chimpanzees also use trial-and-error in making tools. Stripping leaves, peeling bark, splitting stems and biting vines to length requires a mental image and a set of motor patterns, but do not require contingency plans or simultaneous control of more than one variable.

In sum, both specific and general comparisons place the tool-making procedures of the Oldowan well within the abilities seen in modern apes.




Even if Oldowan tools fail to show characteristics that would place them outside the range of ape technology, it is possible that they were used in a more complex way. Because we cannot observe Oldowan tool-use, as we can for apes, we must rely on inference for both specific and general comparisons. In the last few years, the sophistication of archaeological analyses of Oldowan refuse has increased dramatically and the result has been a more reliable, if not uncontroversial, picture of hominid subsistence. This picture appears to corroborate the ape-grade technology documented by the tools (see table 2).

Oldowan tool-makers processed small animals and parts of large animals with their tools. Sites excavated at Olduvai Gorge and at Koobi Fora have stone tools in association with broken-up bones of various sizes and species of animal (Leakey 1971; Isaac 1976). The variety of species and body-parts effectively eliminates the possibility that the bone accumulations resulted from known carnivores. Moreover, some of the bones have cut-marks that seem certain to have been made by stone tools (though note the cautionary findings of Behrensmeyer et al. 1986). Exactly what products the butchers got from the carcasses is the subject of much current debate. Some argue that marrow was the main goal (Binford 1981), others that it was meat (Bunn 1981). Probably it was both (Potts 1984). While it is not yet possible to assess the relative importance of meat and marrow from large (i.e. greater than 100 kilograms body-weight) bovids in the overall Oldowan diet, their frequent presence is enough to distinguish Oldowan foraging from that of chimpanzees.


TABLE 2. Tool-use compared.


Chimpanzee 1

Oldowan 2

1.   What gets processed?

termite mounds, bee hives, ant nests nuts water, brains, honey

1.   What gets processed?

body-parts of large and small animals

2.   How is it processed?

breaking open by pounding and prizing

inserting and luring

guiding and concentrating


2.   How is it processed?

breaking open by pounding and cutting slicing (e.g, tendons)

3.   Where does processing occur?

at termite mounds, ant nests, bee hives

at nut-cracking stations

3.   Where does processing occur?

at death-sites

at processing stations

4.   What gets carried?

raw materials





4.   What gets carried?

raw materials



flaked cores


5.   How far?

nuts-265 m

hammers-500 m

probes-1 km

body-parts-6 km

5.   How far?

at Olduvai

gneiss (rare) 6-13 km

quartzite 2-5 km

lava km

at E. Turkana

from nearby stream bed

1, 2 See Table 1 for references


Chimpanzees also consume small animals and parts of large animals. Apart from eggs and nestlings, the prey are usually other primates and the young of ungulates (McGrew 1983). The largest recorded prey of which parts are eaten is adult bushbuck (Tragelaphus scriptus) which weigh from 25-80 kilograms (Haltenorth & Diller 1977). However, with one notable exception, no ape has been seen to use a tool to obtain meat. The exception involved an old male chimpanzee at Gombe who threw a stone at a defensively massed sounder of bush-pigs (Potamochoerus porcus), hitting one of them. Soon after the pigs broke ranks and in the ensuing melee the chimpanzees captured piglets which they ate (Plooij 1978).

This apparent contrast in tool-use in processing meat deserves further development. Oldowan extractive foragers used stone tools to break into things. Most obvious was their use of tools to break into the diaphyses of long bones in order to get the marrow, but the idea of ‘extraction’ can be extended to butchery as well (Parker & Gibson 1979). The carcass of the prey must be ‘broken into’ in order to get at the edible tissues inside the skin container. Butchery marks on the epiphyses suggest that other products such as tendons may also have been a goal but the many comminuted bones at sites such as FLK suggest that extraction, especially of marrow, was a common procedure in the tool repertoire (Potts 1984). Parker and Gibson were the first to emphasize the evidence for extractive procedures in Oldowan tool-use, but perhaps they inadvertently masked its importance by adding to it several hypothetical behavioral patterns such as aimed throwing.

Chimpanzees also employ extractive procedures in their natural tool-use. Probing for termites, breaking open nuts, and using sponges to swab out fluids or brains are all extractive tasks; all entail removing embedded food from a container. Conceptually this is no different from extracting marrow from bones or meat from carcasses. Use of sponges to extract brains (Teleki 1973) seems indistinguishable from the kinds of procedures that Oldowan meat-eaten used to get at edible portions of carcasses. The behavioral patterns are similar enough to suggest that extractive foraging in the archaeological record can be just as readily interpreted as homologous or analogous. It certainly does not distinguish hominids from pongids.

Much has been written about the significance of the hominid re-use of sites, but again close examination reveals an ape-like pattern. Recent work by Potts (1984; 1986) and Toth (1985) stresses that Oldowan sites such as FLK and KBS[1] were places to which foragers carried tools and food. These were not death sites, so the acquisitors must have carried carcasses or parts of them to these places for further processing or consumption. However, these sites do not appear to have been ‘home-bases’ as some archaeologists maintained initially and enthusiastically (e.g. Isaac 1978). (Evidence of whether the consumers slept there or even spent long periods there would be hard to obtain, however.) By studying weathering patterns on the fossilized bones, Potts (1986) has shown that the bones at FLK accumulated over a period of five to ten years. Over the same period the site was visited repeatedly by carnivores. Also, most of the body-parts taken to FLK were small; few seem large enough to share. The East Turkana site of FxJj 50 (Bunn et al. 1980) consists of several dense scatters of bone associated with stone tools; suggesting several episodes of use spanning a period of not more than a year (authors’ estimate). Like FLK, the site was also visited by carnivores during the period of bone deposition. All this suggests that Oldowan sites such as FLK were picnic-sites (such as groves of trees?) to which foragers carried parts of prey however acquired to be consumed in relative safety and comfort (Potts 1984). Other sites such as the HAS site at Koobi Fora (Isaac et al. 1976) and the ‘Deinotherium’ site at Olduvai (Leakey 1971) appear to have been acquisition sites from which the hominids removed body-parts. In addition to take-away foods, they also carried tools and unmodified stone (Ohel 1984). Some of the stone was carried from as far away as six kilometers, though shorter distances of less than two kilometers were far more common. In sum, Oldowan foraging seems to have included acquisition of parts of prey, which were then carried to another place for further processing with tools, which had also been carried, for consumption.

Chimpanzees practice a directly comparable kind of foraging. They use hammers to crack open nuts and select hammers of  particular sizes and raw materials to fit the task (Sugiyama & Koman 1979; Boesch & Boesch 1983; 1984; Kortlandt & Holzhaus 1987). Because stone is relatively rare in the Tai forest it must be carried to most nut-cracking stations. In most cases processing occurs near the nut trees. However, when suitable anvils or roots or stones are not nearby, then chimpanzees transport nuts as well. In a study of wild-born chimpanzees released from captivity who spontaneously took up nut-cracking, the apes carried nuts several hundred meters to suitable anvils (Hannah & McGrew 1987). However, the extensive data-set from Tai shows that only about 5 per cent. of hammer-transports exceeded 200 meters, and the modal distance was only 5-20 meters (Boesch & Boesch 1983). Chimpanzees re-use nut-cracking stations for years, in some instances producing impressive accumulations of refuse. Finally, chimpanzees sometimes carry away animal food for later consumption: an adult female occasionally nibbled at a piece of bushbuck pelvis which she carried for two days while traveling six kilometers (Hasegawa et al. 1983).

In sum, the spatial pattern of this foraging is the same as can be reconstructed for the Oldowan: carrying tools and food to processing stations that are re-used over time. The specifics differ but the general pattern and its cognitive demands are the same. The important point is that both hominoids carry objects to goals that are out of sight.

Neither the food processed nor the spatial arrangement of processing places Oldowan foraging outside the range of ape behavior. These two aspects of foraging are relatively amenable to study in prehistory, but they do not exhaust the topic of foraging. How did the meat-eaters acquire the body-parts of prey in the first place? There is much debate among archaeologists on this point (e.g. Shipman 1986). If they simply scrounged bones for marrow, then their foraging might have been largely indistinguishable from chimpanzees’ foraging for nuts. If, at the other extreme, Oldowan tool-users killed and butchered large mammals using projectiles and elaborate tool-kits, then their foraging was outside the known range of ape behavior. Few argue for either of these extremes. Most opinion focuses on a flexible meat-getting strategy combining scavenging with hunting as the behavior behind the bone accumulations (Tooby & DeVore 1987). Whether the former was ‘early’ scavenging soon after the prey’s death, Or ‘late’ scavenging of carrion is not yet clear (Bunn 1981; Potts 1984). Whether such hunting involved bringing down healthy adults of large body-size or dispatching injured or aged prey or immatures remains to be determined.

So, how do chimpanzees obtain their meat? Goodall’s (1968) and Teleki’s (1973) descriptions of their hunting (that is, stalking, pursuit, capture, killing) followed by processing (that is, dismembering, sharing, eating) are well-known. Gombe’s apes also show ‘early’ scavenging in pirating most of their bushbuck fawns from the baboons who have just killed them (Morris & Goodall 1977). More recently, Hasegawa et al. (1983) have reported that chimpanzees at Kasoje in Tanzania scavenge ungulates, probably the cached kills of solitary carnivores. Conspicuously absent in all this is the use of tools to obtain meat, except for Plooij’s report (1978) noted above.

To be meat-eaters, early hominoids and living apes must compete with large carnivores; that is, Canidae, Felidae and Hyaenidae. That early hominoids had such competition is clear from the fossil record, but how they coped is yet unknown. For chimpanzees, evidence shows that they can deal with solitary big cats such as leopards (Panthera pardus). This is known from extensive field experimentation (e.g. Kortlandt 1965) and from a striking recent report in which a party of chimpanzees mobbed an adult female leopard and killed her offspring (Hiraiwa-Hasegawa et al. 1986). However well they cope with solitary large carnivores by mobbing, this is a different proposition from dealing with large carnivores in groups; then the apes retreat to the trees (Turin et al. 1981). Thus, lion (Panthera leo), spotted hyena (Crocuta crocuta) and wild dog (Lycaon pictus) probably provide sterner tests. The ape’s solution is one of indirect competition, to be active by day, while the social carnivores are largely crepuscular or nocturnal hunters.

In sum, what we know about Oldowan foraging seems to be within the capabilities of apes. The details differ, but this might be expected given local differences in habitat. Oldowan foraging appears to have been that of a hominoid who lived in a semi-arid, open grassland habitat and who combined scavenging of carcasses and hunting of small game. There is no evidence for a dramatic re-ordering of general hominoid foraging, nor for a evolutionary leap in the cognitive capabilities underlying it.




When we consider the archaeological record of the Oldowan from the perspective of the behavior of modern apes it seems very familiar. The spatial concepts employed to manufacture tools, the natural groupings of tools and even the general procedures for tool-making can all be found in the repertoire of living apes. Moreover, Oldowan foraging patterns such as extractive foraging and the re-use of processing sites are typical of ape patterns of foraging too. The specifics of the Oldowan adaptive niche differ from those of modem apes, but the general pattern is the same. There is nothing about the Oldowan which demands human-like behavior such as language, ritual, shared knowledge of arbitrary design, or other sophisticated material processes. At most one can argue that the Oldowan pushed the limits of ape grade adaptation; it did not exceed them.




The preceding conclusion has relevance for both our understanding of hominoid adaptation and for our understanding of human evolution. In the first case, the analyses re-emphasize the breadth and potential of an ape grade of intelligence and culture. We tend to assume that the range of ape adaptive solutions extends only to those that we see in the modern world, while at the same time we inflate the human range with prehistoric examples. From the fossil record, especially that of the Miocene, we know that hominoids were much more diverse and successful in the past (Pilbeam 1986). This diversity could well have included varieties of technological and foraging patterns not seen in modern pongid populations. We do not disparage Oldowan hominids by placing them in the company of apes; apes were successful for a long time. The use of flaked stone makes the Oldowan look new, but it appears to be a variation on an old theme.

In its general features Oldowan culture was ape, not human. Nowhere in this picture need we posit elements such as language, extensive sharing, division of labor or pair-bonded families, all of which are part of the baggage carried by the term human. The same conclusion applies to the specific elements of behavior which can be reconstructed. However, this does not mean that the Oldowan tells us nothing about the emergence of human behavior. The importance of an ape’s perspective for human evolution is that it allows us to factor out what is not new about the Oldowan and identify what is new. When we factor out the apelike patterns, two others remain that are foreign to modern apes: carrying objects for thousands of meters and competing successfully with large carnivores for large animals. Neither is beyond the competence of an ape adaptive grade, but both point in the direction of advances later in human evolution that eventually clearly distinguish humans from our near cousins.

At first the difference in distances that objects are habitually carried may seem unremarkable: the cognitive prerequisites for carrying objects for 300 meters are presumably no different from those for carrying objects for 3,000 meters. However, the difference is one order of magnitude. This suggests several points. First, hominid bipedalism which frees the upper limbs is much more efficient for carrying objects than is pongid tripedalism or quadrupedalism, as recognized long ago by Hewes (1961). Second, Oldowan foragers had to organize larger units of geography than do modern chimpanzees. The overall home-ranges need not have been larger, but the Oldowan tool-makers clearly planned tool-use that incorporated areas of hundreds of square kilometers. We have no evidence that their spatial repertoires were any different from those of apes, given, for example, the simple geometry of the stone tools, but the management of large geographic spaces may have called for different sorts of spatial abilities from those of apes. Later, soon after the Oldowan, we do find evidence for more complex spatial concepts. Early Acheulean bifaces, 1.5 million years old, required rudimentary ideas of spatial measurement and symmetry (Wynn 1989). Neither concept is known for apes and both suggest a more abstract spatial frame of reference than any used in the Oldowan. Interestingly, neither appears necessary to the mechanical performance of the tools. Perhaps they were borrowed from the repertoire used to organize geographic space, where such abstraction might have been highly selected for. In other words, the relatively complex spatial competence of early Acheulean hominids may have been a response to a situation first encountered in the Oldowan.

Competition with carnivores seems a more daunting problem. If Oldowan meateaters scavenged, then they probably had to employ strategies to deal with large, social carnivores. One possible strategy not often mentioned in the speculative literature is deception. Recent studies of non-human primates suggest that deception in social behavior may have been an important factor in the selection for intelligence (Byrne & Whiten 1988). A shift from its use in within-species competition to across-species competition seems plausible. Moreover, deception probably became a useful tool later in human evolution when full-scale co-operative big-game hunting became prominent. Deception of competing carnivores (e.g. by concealing weapons or by feigning indifference before snatching a prey) would be the kind of behavior that might well be selected later for intelligence of a human grade.

Although we have focused exclusively on the archaeological record in these analyses, the fossil evidence needs mentioning too. If the Oldowan tool-maker was Homo habilis, our conclusion that it had an ape grade adaptation may seem counter-intuitive. After all, the brain of Homo habilis falls between that of living apes and humans in terms of its size and shape (Passingham 1982; Holloway 1981; 1983). This need not mean, however, that Homo habilis must have had a culture that was equally intermediate. If we were to reject archaeological evidence because it was inconsistent with our interpretation of the anatomical evidence, we would come perilously close to circular reasoning: the Oldowan must have been sophisticated because its makers had big brains, therefore culture selected for big brains. (A further cautionary note about the difficulties of assigning tools to purported makers based on anatomical grounds comes from Susman’s (1988) analyses of Paranthropus fossils. He concludes that they, and not sympatric Homo, were the makers.) We do not question the evidence for brains but instead think that scholars must not reconcile the ‘inconsistency’ by promoting the Oldowan to insupportable cultural levels. The evolutionary pressures for encephalisation may have been unrelated to culture. If one uses archaeological evidence, it must speak for itself, and the Oldowan argues for an ape adaptive grade.




An earlier version of this article was delivered by Wynn at a conference, ‘Tools compared: the material of culture’, sponsored by the Royal Anthropological Institute, London, 1988.





Behrensmeyer, A.K., K.D. Gordon & G.T. Yanagi 1986. Trampling as a cause of bone surface damage and pseudo-cutmarks. Nature. Lend. 319, 768-71.

Bernstein, I.S. 1962. Response to nesting materials of wild born and captive born chimpanzees. Animal Behav. 10, 1-6.

Binford, L.R. 1981. Bones: ancient men and modern myths. New York: Academic Press.

Boesch, C. &, H. Boesch 1983. Optimisation of nut-cracking with natural hammers by wild chimpanzees. Behaviour 83 , 265-86.

——— 1984. Mental map in wild chimpanzees: an analysis of hammer transports for nut cracking. Primates 25, 160-70.

Boysen, S.T., C.G. Berntson & J. Prentice 1987, Simian scribbles: a reappraisal of drawing in the chimpanzee (Pan troglodytes). J. Comp. Psychol. 101, 82-9.

Bunn, H.T. 1981. Archaeological evidence for meat-eating by Plio-Pleistocene hominids from Koobi Fora and Olduvai Gorge. Nature, Land, 291, 574-80.

——— , J.W.K. Harris, G. Isaac, Z. Kaufulu, E. Kroll, K. Schick, N. Toth & A. Behrensmeyer 1980. Fxjj50: an early Pleistocene site in northern Kenya. WId. Archaeol. 12(2), 109-39.

Byrne R.W. & A. Whiten (eds) 1988. Machiavellian intelligence: social expertise and the evolution of intellect in monkeys, apes, and humans. Oxford: Univ. Press.

Chiarelli, B. & R.S. Corruccini (eds) 1981. Primate behaviour and sociobiology. Berlin: Springer.

Coppens, Y. F.C. Howell, G. Isaac & R. Leakey (eds) 1976. Earliest man and environments in the Lake Rudolph Basin. Chicago: Univ. Press.

Foley, R. (ed.) 1984. Hominid evolution and community ecology: prehistoric human adaptation in biological perspective. London: Academic Press.

Goodall, J.L. 1968. The behaviour of free-living chimpanzees in the Gombe Stream Reserve. Animal Behav. Monogr. 1, 161-311.

Gowlett, J. 1984. Mental abilities of early man: a look at some hard evidence. In Foley 1984.

Haltenorth, T. & H. Diner 1977. A field guide to the mammals of Africa including Madagascar. London: Collins.

Hannah, A.C. & W.C. McGrew 1987. Chimpanzees using stones to crack open oil palm nuts in Liberia. Primates 28, 31-46.

Harris, J.W.K. 1983. Cultural beginnings: Plio-Pleistocene archaeological occurrences from the Afar, Ethiopia. Afr. Archaeol. Rev. 1, 3-31.

Hasegawa, T., Hairaiwa, M., Nishida, T. & H. Takasaki 1983. New evidence on scavenging behaviour in wild chimpanzees. Curr. Anthrop. 24, 231-32.

Hewes, G.W. 1961. Food transport and the origin of hominid bipedalism. Am. Anthrop. 63, 687-710.

Hiraiwa-Hasegawa, M., Byrne, R.W., Takasaki, H. & J.M.E. Byrne 1986. Aggression towards large carnivores by wild chimpanzees of Mahale Mountains National Park, Tanzania. Folia primat. 47, 8-13.

Holloway. R.L. 1981, Exploring the dorsal surface of hominoid endocasts by stereoplotter and discriminant analysis. Phil, Trans. R. Soc. Lond. B292, 385-94.

——— 1983. Human brain evolution: a search for units, models and synthesis. Cana. J. Anthrop. 3, 215-30.

Isaac, G.L. 1976. Plio-Picistocene artifact assemblages from East Rudolph, Kenya. In Coppens et al. 1976.

——— 1978. The food-sharing behavior of protohuman hominids. Sci. Am. 238(4), 90-108.

——— 1981. Archaeological tests of alternative models of early hominid behaviour: excavation and experiments. Phil. Trans. R. Soc. Lond. B292, 177-88.

——— 1984 The archaeology of human origins: studies of the lower Pleistocene in East Africa 1971-1981. Adv. Mld Archaeol. 3, 1-86.

Kitahara-Frisch, J., Norikoshi. K. & K. Hara 1987. Use of a bone fragment as a step towards secondary tool use in captive chimpanzee.  Primate Rep. 18, 33-7.

Kortlandt, A. 1965. How do chimpanzees use weapons when fighting leopards? Yb. Am. phil. Soc. 1965,

Kortlandt, A. & E. Holzhaus 1987. New data on the use of stone tools by chimpanzees in Guinea and Liberia.
Primates 28, 473-96.        Leakey, M. 1971. Olduvai Gorge, vol. 3. Cambridge: Univ. Press.
McGrew, W.C. 1974. Tool use by wild chimpanzees in feeding upon driver ants. J. hum. Evol. 3, 501-08.
1983. Animal foods in the diets ofwild chimpanzees (Pan troglodytes): why cross-cultural variation?
            J. Ethol. 1, 46-61.
McGrew, W. C. & D.A. Collins 1985. Tool use by wild chimpanzees (Pan troglodytes) to obtain termites (Macrotermes
herus) in the Mahale             Mountains, Tanzania. Am. J. Primat. 9, 47-62.
Morris, K. & J. Goodall 1977. Competition for meat between chimpanzees and baboons of the Gombe National Park.
            Folia primat.
28, 109-2 1.
Ohel, M. 1984. Spatial management ofhominid groups at Olduvai: a preliminary exercise. Palaeoecol. Africa 14,125-46.
Parker, S.T. & K.R. Gibson 1979. A developmental model for the evolution oflanguage and intelligence in early hominids.
            Behav. Brain Sci.
2, 367-408.
Passingham, R.E. 1982. The human primate. Oxford: W.H. Freeman.
Piaget, J. 1972. The Principles of Genetic Epistomology (trans.) W. Mays, London: Kegan Paul.
Pillbeam, D. 1986. Distinguished lecture: horninoid evolution and horinnoid origins. Am. Anthrop. 88, 295-312.
Plooij, FX 1978. Tool-use during chimpanzees' bushpig hunt. Carnivore 1(2), 103-06.
Potts, R. 1984. Hominid hunters? Problems of identifying the earliest hunter/gatherers. In Foley 1984.
_______1986. Temporal span of bone accumulations at Olduvai Gorge and implications for early hommid
foraging behavior. Paleobiol. 12, 25-31.
Sabater, Pi, J. 1974. An elementary industry of the chimpanzees in the Okorobiko mountains, Rio Muni
(Republic ofEquatorial Guinea), West Africa. Primates 15, 351-64.
Shipman, P. 1986. Scavenging or hunting in early honunicls: theoretical framework and tests. Am. Anthrop. 88,27-43.
Smith, D.A. 1973. Systematic study of chimpanzee drawing. J. compar. physiol. Psych. 82, 406-14.
Sugiyama, Y. 1985. The brush-stick of chimpanzees found in south-west Cameroon and their cultural characteristics. Primates 26, 361-74.
_______& J. Koman 1979. Tool-using and -making behavior in wild chimpanzees at Bossou, Guinea.
Primates 20, 513-24.
Susman, R.L. 1988. Hand of Paranthropus robustus from Member 1, Swartkrans: fossil evidence for tool behavior. Science 240, 781-84. Tanner, N.M. 1987. The chimpanzee model revisited and the gathering hypothesis. In The evolution of human behaviour: primate models.               (ed.) W.G. Kinzey. Albany: State Univ. of New York Press.
Teleki, G. 1973. The predatory behavior ofaild chimpanzees. Lewisburg: Bucknell Univ. Press.
Tobias, P.V. 1983. Hommid evolution in Africa. Can. J. Anthrop. 3, 163-85.
Tooby, J. & 1. DeVore 1987. The reconstruction of honunid behavioral evolution through strategic modelling. In The evolution of human                behavior: primate models. (ed.) W.G. Kinzey. Albany: State Univ. of New York Press.
Toth, N. 1985. The 01dowan reassessed: A closer look at early stone artifacts. J. Archaeol. Sci. 12, 101-120.
Turin, C.E.G., W.C. McGrew & Pj. Baldwin 1981. Responses ofwild chimpanzees to potential predators. In Chiarelli & Corruccini 1981.
Wright, R. 1972. Imitative learning of a flaked stone technology-the case of an orangutan. Mankind 8, 296-306.
Wynn, T. 1978. Tool-using and tool-making. Man (N. S.) 13, 137-8.
________1981. The intelligence ofOldowan hominids. J. Hum. Evol. 10, 529-41.
________1989. The evolution of spatial competence. Champaign: Univ. of Illinois Press.


[1]     FLK is the abbreviation established by Louis Lcakey for Frieda Leakey Korongo, one of the archaeological localities in Bed I at Olduvai Gorge. KBS and HAS are abbreviations for specific Oldowan sites at Koobi Fora in northern Kenya. Fxjj50 is the designation for another Oldowan site at Koobi Fora, in this case labelled by a formal site enumeration system used throughout Africa.