There is an inherent problem within African Acheulian and Middle Stone Age assemblages research: how to recognize culture and ethnicity and the period in which hominins evolved the cognitive ability to invent social solutions to perceived or real ecological and functional challenges. There also exists the challenge of establishing the relationships between the stone tool assemblages and the cultures of which the knappers were a part.
These problems are investigated through functional, compositional, ecological, and temporal analyses of the Acheulian handaxes, the Acheulian-Middle Stone Age transition, the timing of the occupancy of the central African rain forests, the Still Bay level at Blombos Cave, and the Middle Stone Age layers at Klasies River. African archaeologists have used the typological categories introduced by Goodwin and Van Riet Lowe and the lithic Modes of Clark to create a conceptual framework in which they could operate; however, the framework has produced tensions and constraints whose chronological, cognitive, cultural, functional, and environmental problems have become prevalent in the Middle and Late Pleistocene research paradigms of Africa.
The development of more sophisticated models is required on an integrated regional basis in order to capture and subsequently extrapolate the differential expressions of lithic technology, temporally and spatially.
John Goodwin and Clarence van Riet Lowe published a seminal work in 1929 which has influenced African archaeology ever since. Entitled Stone Age Cultures of South Africa, Goodwin and Van Riet Lowe proposed that the Stone Age sequence be subdivided into the Early, Middle, and Late Stone Age (hereafter shortened to ESA, MSA, and LSA). By proposing their alternative scheme to the European sequence of Lower, Middle, and Upper Palaeolithic, Goodwin and Van Riet Lowe were emphasizing the distinctiveness of the African archaeological temporal sequences and typological assemblages.
The post-World War II period witnessed the beginnings of a major overhaul of the archaeological discipline. Beginning with the advent of radiocarbon dating in the 1950s and continuing with the New Archaeology’s hypothetico-deductive-nomological model, archaeology drew upon and attempted to reorient itself within the hard sciences. Goodwin and Van Riet Lowe proposed a classificatory scheme that was formally approved by the Third Pan-African Congress held in 1955.
The classificatory schemes outlined above for Africa and Europe were taken a step further and recombined by J. D. Clark on the basis of dominant lithic technologies. His resulting Modes of Technology divide the history of stone tools into five modes. The modes are held to be reflective of raw material availability, functional differentiation and manifestations of hominin technological strategies. The table outlines the current status of lithic mode recognition and classification.
Technological Mode Industry and Fossiles Directeurs
Mode 1 Oldowan, Early Stone Age (choppers and flakes)
Mode 2 Acheulian, Early Stone Age (bifacial hand-axes)
Mode 3 Middle Paleolithic, Middle Stone Age
(prepared cores, points)
Mode 4 Upper Paleolithic (retouched blades)
Mode 5 Mesolithic and Late Stone Age (microlithic
composite flakes and blades)
After Clark, 1977, and Foley & Lahr, 1997.
These technological strategies are manifestations of behavioral adaptations whereby knowledge and culture are transmitted through social learning. The accumulated repertoire limits the risks of invention in technological and cultural evolution, but permits their expression in a wide diversity of situations through socially mediated responses to particular internal or external stimuli.
The mode classificatory scheme was built on the assumption of inbuilt progress in the evolution of stone tool technology. In this view, progress, combined with varying systems of inheritance, was enough to explain the contrasting time spans of the different lithic technologies. This framework was the intellectual birth of the model seeking to explain the advent of “modern human behavior” as a revolution occurring between 50 kya and 40 kya. It became clear there were serious inherent defects. One such problem and its implications for African stone tool research involved the transitional boundary between the Acheulian and the MSA.
Interwoven with the investigations on the transitional period from the Acheulian to the MSA is the issue of how the stone tools were manufactured, what materials were used, the sources of the raw material procurements and what is signified in terms of the knappers’ functional and cognitive capacities.
Inquiries into the uniformity or regional distinction of Acheulian handaxes need to account for the composition of stone procured, the flaking properties of the raw materials and the reduction sequences used. A 1990 study by Thomas Wynn and Forrest Tierson looked at Middle Pleistocene samples gathered from four regions of the world, including East Africa. The East African sites of Olorgesaille, Isimila, Kariandusi, and Lewa were selected. The results obtained in particular from Kariandusi reinforce previous studies that the form of the Acheulian bifacial implements was constrained by the raw materials utilized. Through function analysis, Wynn and Tierson were able to achieve a 75% success rate in sorting the Kariandusi handaxes into their correct raw material constituents. The remaining unaccounted for 25% was attributed to conscious cultural expression. However, given the time span of the East African sites chosen for the study (700-200 kya), Wynn and Tierson failed to adequately account for the possibility of temporal variation as viable alternative explanation.
This is an argument that has been picked up by Iain Davidson, in particular. Davidson rejects the hypothesis that the Acheulian handaxes were a product of deliberate intentionality. He argues that the observed patterning is insufficient evidence of the intent of the knappers and demonstrates the cognitive inability to communicate using symbols. Defined as “the assumption that the objects which the archaeologist recovers represent final forms corresponding to the self-consciously articulated, prior intentions of their one-time makers,” the Davidson’s Finished Artifact Fallacy hypothesizes differences between artifacts which are distinctive, repetitive and standardized, and those artifacts which have had form imposed upon them. In place of the Finished Artifact Fallacy, Davidson advocates the re-examination of the Acheulian industry due to archaeologists having conducted selective studies through choosing which materials to analyze. He argues that archaeologists have unwittingly introduced patterns based upon both their selections and the procedural practices and technological steps used by the makers of the artifacts. This makes it extremely difficult to arrive at an understanding of the intentions and cognitive capabilities of the makers without eliminating and understanding these biases and steps. Davidson argues that the Acheulian handaxes are examples of this, as it is impossible to distinguish between flakes originating from cores or handaxes, and that this is therefore an indicator to the lack of guiding principles on the part of the knappers.
The functional hypothesis that the type of raw material chosen determined the shape and size of the Acheulian handaxe remains a persuasive argument. Reinforcing this hypothesis is the experimental archaeology showing resharpening reduction is capable, through three to four episodes, of altering an Acheulian handaxe into a Developed Oldowan handaxe. However, this demonstration of a plausible two-stage life suffers from the same inherent flaw as the Wynn and Tierson study: conflation of temporal variation.
Only two of the several clusters of late Acheulian sites mapped by Garth Sampson at Seacow Valley, South Africa, ranged within 1 kilometer of their corresponding quarry. The vast majority of the sites were situated near river beds and still within easy access distance to the quarries. The artifacts show three developments over the symmetrical bifaces of the early Acheulian. The developments have been categorized as congruency, three-dimensional symmetries and broken symmetry. An example of the latter is the twisted profile handaxes that first appear ca. 350 kya. Wynn poses the question as to why symmetry would have been imposed and queries whether it might be connected with mate selection, a hypothesis also advocated by Kohn and Mithen but which ultimately rests on an untestable psychological basis.
Lithic assemblages have often been held up as representative of cultural entities, through either in situ cultural differentiation or population movements into new raw material niches permitting expansion of existing or development of new repertoires. Illustrative of this is the distribution pattern across the landscape of the Lupemban industry, dated to ca. 270 kya, particularly in relation to the prior Acheulian and Sangoan site patterns.
Increasing evidence has been put forward by Lawrence Barham for the Lupemban industry exhibiting a cognitive, technological, and adaptive sophistication not witnessed previously. Using environmental reconstructions, it is claimed that while the Sangoan exhibits a resemblance in distributional patterns to the Acheulian, there was an important difference which contributed to the emerging pattern that was to become the Lupemban industry: the Sangoan sites are generally found in habitats of higher productivity. This hypothesis holds that the loosening of ties to waterways and lakes permitted the subsequent knappers of the Lupemban industry to expand into and inhabit the Congo basin which, in the paleoenvironmental reconstruction, was forested.
Barham builds on the work done at Kalambo Falls, and his own work at Mumbwa Caves and Twin Rivers, to plot the known distribution of Lupemban sites over maps of the paleoenvironmental reconstruction of central Africa. The resultant model takes into account the glacial/interglacial cycles of African biogeography as well as lithic assemblage distribution mappings of 36 years ago. The problems within the model are manifest and raise serious concerns over both its accuracy and the conclusions drawn from it, namely that the central African rain forest was occupied at the time of the Lupemban.
Although climatic fluctuations can be deduced from the Congo fan deposits, these deposits do not extend beyond 200 kya. Neither do there exist reconstructions of climatic sequences from the internal
Congo basin itself. While it is claimed that Lupemban sites existed in closed forests in Equatorial Guinea, through marine pollen profiles, more precise analyses are required for local areas and from the sites in question.
Barham claims that the Lupemban industry reflects a new composite stone tool technology: backed blades and segments, bifacial lanceolates, notched bifacial points and tranchets. These, according to his hypothesis, are the tools of a specialized hunting and gathering culture with the technological and social capabilities of extracting the nutritional requirements from closed canopy plant and animal resources.
Although Barham recognises the economic arguments brought to bear against an early occupation of the rain forests of central Africa, he fails to adequately address the concerns. The nutritional values of plant and animal resources in closed canopy forests are such that it appears unlikely modern hunter-gatherers occupied closed canopy forested areas for substantial periods of time prior to the latter portion of the LSA. If this turns out to be correct, then serious questions have to be posed about the viability of hominins occupying and surviving in these environments 270 kya.
Alternative hypotheses are required to be formulated which do not relate the Lupemban composite assemblage to a new cultural, and possibly ethnic, grouping adapted to survival in forested and semi-forested environments. More profitable lines of inquiry may look at how the Lupemban was adapted to savannah and semi-woodland areas, the variability in intersite assemblages and how the composition of this expression could be compared culturally and technologically with contemporary and later variable compositions.
Still Bay sites are only found in sub-Saharan Africa and are predominantly along the East Cape south coast. The dating of the Still Bay industry has long been an unanswered question in southern African archaeology. The results of radiocarbon dating on charcoal and shell from the Still Bay levels at Blombos Cave, on the southern Cape coast of South Africa, should be considered as infinite dates. Recent luminescence dates suggest an age of ca. 100 kya for the earliest MSA layer, BBC 3. The Still Bay points come from the BBC 1 phase which underlies a sand dune layer dated by optically stimulated luminescence to ca. 70 kya.
The diagnostic feature of the Still Bay industry is the leaf-shaped, bifacially retouched point possibly used as a spearhead. The primary raw material used was silcrete. Length can vary from two to three centimeters to the very large “Blombos point.” Deacon has suggested that this assemblage is indicative of craft specialization, because its function as a hunting spear point could possibly have been better served had the points been smaller. This hypothesis therefore accords stylistic symbolism to the artifacts.
Although there are no Mode 3 occurrences in Africa, and thus no African antecedent for the European Upper Paleolithic, certain viewpoints grew up around the question of the perceived lack of variation and change in the MSA. These are problematic from the viewpoint of the temporality and compositional significance of the Howiesons Poort industry in southern Africa. No site has been more central to the development of the debate than Klasies River.
Klasies River is situated on the southern-east coast of South Africa. The site is comprised of five MSA and one Holocene deposits from caves 1,1A, 1B, 1C, and 2.
Oxygen isotopic analyses and uranium disequilibrium dating place the basal layer in correlation with MIS (Marine Isotope Stage) 5e. It has been proposed that the SAS member (MSA I and II) is in the region of 90 kya and the Upper Member (MSA III and IV) no less than 50 kya, thus out of the range of radiocarbon techniques. Contra Parkington, who hypothesized that southern African Howiesons Poort assemblages may also postdate 50 kya, the Howiesons Poort layer is underlain and overlaid by MSA II and III respectively. The Howiesons Poort is generally correlated with MIS 5a and 4, the boundary of which is estimated at around 74 kya.
Retouched blades were Clark’s fossile directeur for Mode 3. Yet although blades are one of the main characteristics of the MSA and backed artifacts of the Howiesons Poort, neither should be viewed as anticipating the advent of a lithic assemblage on a separate continent. In the case of the Howiesons Poort, there is up to 25 kya in time between it and the Aurignacian.
The correlation of the Klasies River MSA sequences with periods of changing environmental conditions and the changing assemblage compositions led Ambrose and Lorenz to propose an ecological deterministic model of changing social conditions. Through analyses of resource structure, they claimed that the MSA inhabitants failed to respond to the changing environmental conditions in the same flexible patterns as their Holocene counterparts. These results indicated that their socioterritorial and technological systems were determined solely by the resource ranges available for exploitation with the least energy output.
While recognizing that the environmental stress caused by ecological changes at the start of the Howiesons Poort led to the increased resource ranges postulated by Ambrose and Lorenz, their hypothesis fails to explain why access to new raw materials would automatically result in the advent of new stone tool artifacts. In order for their functionalist ecological hypothesis to account for these new artifacts, Ambrose and Lorenz would be required to “ascertain for every concrete case the causal chain which linked the initial natural impetus and the resultant cultural transformation.”
Hilary Deacon has proposed a structuralist alternative whereby the environmental stresses placed on the Howiesons Poort occupants resulted in attempts to maintain previously existing social structures and population levels: Instead of adapting to nature, they modified their behavior. Remnants of their adaptation can be found through examination of their stone artifacts in functional terms. He concludes, through analogies with backed artifacts from the southern African Wilton assemblages (8—4 kya), that the backed segments and trapezes are best regarded as hafted projectile points used in hunting; this analogy has subsequently been weakened by microwear and experimental analyses suggestive of the Wilton segments being components of knives instead of functioning as arrowheads. The MSA backed pieces are larger and were more likely to have been spear points and barbs. Ethnographic writings on the Bushmen record that projectile points carry both stylistic and social meaning, marking social and linguist groups, and are used as exchange gifts. The appearance of backed artifacts in the Howiesons Poort is, therefore, viewed as an attempt by the toolmakers to mark out social boundaries and the intensification of social networks under a period of environmental duress. As the climatic conditions improved toward the onset of the MSA III, the factors that influenced this cultural selection for a high level of symbolic behavior therefore gradually fell away.
The challenge posed by Deacon’s generalized hypothesis was taken up by Sarah Wurz who, in 1995, took a random surface sample of Howiesons Poort artifacts from Cave 2. Comparisons were made with the 1968 Cave 1A top cutting (119 336 artifacts) sample (KR1A-68) of Singer and Wymer by using the use of the same major typological categories: waste, used pieces or edge-damaged pieces, and formal artifacts.
Lindly and Clark theorize that the continuous lithic industries reveal no evidence of stylistic patterning over time; they concede that the Howiesons Poort layers (Layers 10-21 of Shelter 1A) demonstrate a shift in the MSA to retouched stone tools. However, they make no reference to a recent technological and typological analysis of the MSA II stage at Klasies River Mouth, conducted by Thackeray and Kelly, that concludes the finished product does not necessarily represent they way things are in practice, but rather the way things should be ideally.
Howiesons Poort blades occur in greater frequency than their typical MSA predecessors and descendants, 17.44% versus 0.98%. There is also a lack of retouched and used artifacts. The backed artifacts have previously been suggested by Thackeray to be less standardized than their Later Stone Age counterparts, a conclusion countered by Wurz. The variation coefficient obtained falls within that recorded for the LSA, demonstrating the backed artifacts from the Howiesons Poort and the Later Stone Age were designed with a comparable mental construct of their shape parameters. Backed artifacts from the MSA Howiesons Poort and the Later Stone Age Wilton are design types of the same kind, although the Howiesons Poort backed artifacts range from two to three times their size.
Strengthening Deacon’s argument, Wurz draws parallel between demonstrable archaeological and ethnographic usage of backed artifacts as projectiles. Amongst the Bushmen these projectiles are imbued with symbolism. Projectiles can take many possible forms and therefore the form chosen during the Howiesons Poort was a material action of choice governed by social rules and by extension was a form of modern behavior.
By comparing the raw material samples of KR2-1995 and KR1A-68 in terms of quartzite and nonlocal rock occurrences, Wurz determined that a marked difference is evident in the raw material composition. Around 27% of its Howiesons Poort artifacts were manufactured from non-quartzite (and consequently nonlocal) raw materials. By contrast the unbiased KR2-95 is comprised of 4% non-quartzite artifacts, with a high percentage (39%) of the retouched artifacts made from these materials. A high degree of selection is therefore present in the choice of raw materials.
Wurz used the analytical chaine operatoire approach in analyzing the choice of raw materials by the Klasies River people for their stone tools. Nonquartzite materials increase in numbers only in the middle stages of the Howiesons Poort, before again falling to low levels. Wurz attributes these raw material changes to conscious choices.
Wurz analyzed the backed artifacts in terms of segments (crescents), intermediates and trapezes to determine whether a relationship could be established between the form and extent of backing, and the resulting artifact. The waste materials are also regarded as by-products from the main design of shaping cores for the removal of blanks. Consequently, the production and subsequent use of backed artifacts were likely to have been the result of decision making and deliberate steps taken.
The sites of Kapthurin and Herto exemplify warnings about using typology uncritically as chronology. The mixture of the typologies in the lithic assemblages over a period of time in excess of 100 kya is a timely reminder that a direct correlation between typology and a set chronological timeframe is an uneasy assertion. The mixing of the Herto assemblage is on a greater scale than has previously been envisaged. Kapthurin poses a particular problem to mode model formulations, as the assumptions of progress inherent within do not take into account the mix of technological strategies that were employed.
It is therefore questionable whether the onset of Mode 2 is indicative of hominins having developed the ability to impose symmetry upon a stone tool or whether it was an extension of the process of tool reduction developing out of the Oldowan, Mode 1. It would be unwise to postulate a single purpose and function for Acheulian bifaces that flourished and evolved over 1.5 million years, and single purpose ideas such as sexual selection are untestable. It is more likely that the current opposing models of symmetry and reduction sequences are disguising real temporal and productive variations, which require greater chronological and stratigraphic control over the source sites to elucidate. The sequences involved are complex and varied and what is visible is the byproduct of that variability, variability that itself might have changed over the course of hundreds of thousands of years. This changing variability is most vivid in the late Acheulian within the transitional period to the MSA, with Acheulian handaxes being knapped by the Levallois method.
The variability of the new MSA assemblages has led to hypotheses over cultural markings and new population settlement patterns. A recent hypothesis by Barham regarding the occupation of closed canopy forests in Central Africa is questionable on the grounds of nutritional deficiencies of rain forest resources and on the low level of climatic resolution available. Questionable too is Lahr and Foley’s linkage of the appearance of the new technology with the advent of a new Homo species, Homo helmei, thus linking anatomy with technology and culture.
The Klasies River MSA I and II assemblages date to the last interglacial but, contra the expectations of the Ambrose and Lorenz, there are important differences in their composition. MSA I follows a blade production strategy, with the MSA II a Levallois-like point manufacture. These lead into the Howiesons Poort via a tendency toward smaller blades in the upper MSA II sequence. This cuts across Ambrose and Lorenz’s simplistic technological-ecological indicator models, undermines Parkington’s use of the Howiesons Poort as a transitional phase to the LSA, negates Singer and Wymer’s use of the Howiesons Poort for the arrival of new ethnic inhabitants and is a powerful indicator of changing cultural traditions which are reflected in the compositional makeup of the assemblages.
Ambrose and Lorenz see the Howiesons Poort occurrence as an adaptive response to environmental pressures that resulted in mobility pattern changes and the procurement of higher-quality raw materials for their stone tools. This viewpoint therefore presupposes that the technological changes are not related to stylistic preferences and expressions. To say that the environment determined their responses ignores the mental processes behind the conscious choices involved in making the changes they did. The Klasies inhabitants, for example, had perfectly good quartzite raw material available on the beaches, yet during the Howiesons Poort they imported silcrete from 20 kilometers away to manufacture the backed tools. The manufacture of the Howiesons Poort artifacts was, therefore, more likely to have been governed by social rules.
Deacon has proposed that the Still Bay is an example of a MSA craft specialization industry. Blombos and Klasies River are not a large distance apart, which suggests that the Klasies River inhabitants did indeed have access to such knowledge but made a cultural choice not to use it. Deacon suggests that craft specialization resulted in reciprocal trading, but it still remains to be determined first whether the Still Bay industry is indeed a form of craft specialization. What is clear is that the Still Bay industry points do not make for better spear points than other smaller bifacial tools, and this may have had something to do with a symbolic value being attached to them.
From before, but especially after, the publication of Stone Age Cultures in South Africa, African archaeologists have been increasingly concerned with typology. Cemented in Clark’s lithic Modes, these typologies fostered a belief in a progressive lineage of lithic assemblages in Africa and contributed in particular to the Howiesons Poort being viewed as a transitional phenomenon to the LSA.
The Acheulian handaxes have been used as settlement markers in relation to their quarries and indicative of riverine resources at Seacow Valley. Kapthurin and Herto have mixed Acheulian and MSA industries, and together span of ca. 150,000 years which also covers the period in which the Lupemban is probably inaccurately claimed to have penetrated the central African forests. The changing nature of the manufacturing process of the Acheulian handaxes, the composition of the Lupemban industry and the changing compositional nature of the MSA industries at Klasies River are suggestive of evolving cognitive abilities that are most prominently expressed in the Howiesons Poort. Although environmental factors would have influenced the territorial ranges of the MSA peoples, conscious choices were made as to raw materials for their stone tools.
The untidy transition from the Acheulian to the MSA, the appearance of the regional Still Bay industry at Blombos and the integrity of the MSA at Klasies River, are strong arguments against simplistic lithic mode and cognitive-environmental proposals. The innovation in the Still Bay and the Howiesons Poort, while maintaining lithic patterned links to the preceding and later MSA assemblages, demonstrates the folly of linking new lithic expressions to incoming migrants possessing a different culture; this is paralleled in the caution over linking the Lupemban to new cultural adaptations by a new Homo species on the basis of weak data.
Sophisticated regional models are required, taking into account the full range of technological, faunal, geological, and taphonomic processes. These new models must incorporate variable cultural expressions, constrained adaptations to changing environmental conditions in terms of the existing cultural and social repertoire of the inhabitants, and the desire to link form to function when such linkages should be inferred from the data rather than attributed to it and thus constrain it technologically and temporally.
- Barham, L. (2001). Central Africa and the emergence of regional identity in the Middle Pleistocene. In L. Barham & K. Robson-Brown (Eds.), Human roots: Africa and Asia in the Middle Pleistocene (pp. 65-80). Bristol: Western Academic & Specialist Press.
- Clark, J. D., Beyene, Y., WoldeGabriel, G., Hart, W., Renne, P., Gilbert, H., Defleur, A., Suwa, G., Katoh, S., Ludwig, K., Boisserie, J.-R., Asfaw, B., & White, T. (2003). Stratigraphic, chronological and behavioral contexts of Pleistocene Homo sapiens from Middle Awash, Ethiopia. Nature, 423, 747-752.
- Davidson, I. (2002). The finished artifact fallacy: Acheulian hand-axes and language origins. In A. Wray (Ed.), Transitions to language (pp. 180-203). Oxford: Oxford University Press.
- Fitzhugh, B. (2001). Risk and invention in human technological evolution. Journal ofAnthropological Archaeology, 20, 125-167.
- Henrich, J., & McElreath, R. (2003). The evolution of cultural evolution. Evolutionary Anthropology, 12, 123-135.
- Henshilwood, C., & Marean, C. (2003). The origin of modern human behavior: Critique of the models and their test implications. Current Anthropology, 44, 627-651.
- McPherron, S. (2000). Handaxes as a measure of the mental capabilities of early hominids. Journal of Archaeological Science, 27, 655-663.
- Stringer, C. (2002). Modern human origins: Progress and prospects. Philosophical Transactions ofthe Royal Society of London B, 357, 563-579.
- Wynn, T. (2002). Archaeology and cognitive evolution. Behavioral and Brain Sciences, 25, 389-438.