Only certain kinds of stone can be used for knapping stone tools. It needs to fracture easily and predictably. The smoother the better, which is why the most elaborately knapped material that archaeologists have found is the volcanic glass called obsidian. The slightly less fine-grained and less homogeneous flint and quartz are more widely available in nature and were the main materials used in prehistoric (and historic) times for knapped stone tools. More coarse-grained materials have also been used. We tend to associate these with the more simple stone tools made by the earliest of our ancestors, but they were used throughout prehistory. In areas where good raw materials were not readily available our ancestors have even knapped seashells, limestone or wood into tools.
If the knapper knows how to control the angle and force of the blow and the exact point where the hammerstone hits the rock, he/she can control the shape and size of the pieces that come off the lump of rock (the flakes) and the shape of what remains of the lump of rock (the core). By controlling the shape of both the flakes and the core, the knapper can string together a sequence of strikes to the core to remove flakes, which results in a complex end product.
Humans first used hard hammers (hammerstones) for knapping. They started to use soft hammers (antler, wood or bone) perhaps 700,000 years ago for finer control. Pressure flaking, where soft hammers are pushed (not struck) to cause a fracture, came along in the Upper Palaeolithic, after the time of the Neanderthals.
The flakes or blades that come off the core are sometimes reworked (or, in lithic terminology, ‘retouched’). In this reworking, tiny flakes are knapped off a flake’s edges in order to give it a desired shape (such as straight or curved) or a particular morphology (such as serrated or pointed).
Retouching can also be used to resharpen tools that have become dull from use. In the course of their useful lives, lithic artifacts can be potentially reworked into different forms with different functions. Unlike in most other technologies, a lithic artifact is essentially never finished.
Traditionally a ‘tool’ was considered to be either the desired end product of knapping or a retouched flake. In the last few decades, new techniques to examine microscopic wear marks and residues left on the surface of used stone artifacts have demonstrated that flakes that were not retouched at all were also used regularly. In fact, when a flake is struck off the core its edges are fresh and very sharp (often sharper and thinner than a modern surgical scalpel).
How can archaeologists recognize lithic artifacts? In knapping, when the hammerstone hits the rock, force spreads through the material in concentric shock waves. These shock waves are well pronounced and close to one another near the point of impact and become less marked the further away they travel from the point of impact. Think of the ripples created when a pebble hits the surface of a pond. If the knapped lump of rock splits apart in two pieces, each of these two pieces will have a new, slightly curved surface with concentric ripple marks reminiscent of the ripples of a shell. (Hence the formal name for this type of fracture, ‘conchoidal’, derived from the Greek word for shell.) These two rippled surfaces are each other’s perfect negative. At the point where the hammerstone hits the rock, the flake has a half-conical bulge, which is called the bulb of percussion, while the same spot on the core retains the negative of the bulb.
This ‘conchoidal’ type of fracture, with the bulb and ripple marks that it leaves on the knapped pieces of stone, is the basis of the archaeological analysis of stone tools. At the most elementary level, it provides a way to distinguish stone tools (rocks intentionally knapped by humans) from similar-looking rocks fractured by natural processes such as rockslides. This is particularly useful in early sites, where the tools found are often just simple flakes and may not be accompanied by any hominin fossils or other corroborating finds. Indeed, there are often controversies over the human role in shaping the stones at such sites.
Flint-knapping is essentially a process of peeling material off a lump of rock, layer by layer, much like peeling an onion. Based on the scars and the sets of ripple marks left on a core, it is possible to reconstruct how each successive ‘layer’ of material was ‘peeled’ off the core.
The back of each flake also bears the partial scars left on the core by earlier flake removals. To the lithics specialist, each flake and each core tells a little segmented story of how it was made. Putting together all the little stories from the tens or even thousands of stone tools recovered from a site, one can reconstruct the bigger story of how the stone tools of that site were made.
As the flint-knapper works down a core, removing flakes layer by layer, each piece that falls off the core remains on the ground, around where the knapper sits. In some cases it is even possible to reconstruct where a knapper was positioned at a site. Often the two rippled surfaces can be fitted back together (or ‘refitted’ in lithic analysis lingo). Refit all or most of the flakes removed from the same core and you end up with the original lump of rock in a three-dimensional jigsaw puzzle. This has been done successfully at many sites, notably Boxgrove and Maastricht-Belvédère. It is no surprise that of the two authors of this book, the lithics specialist is also a jigsaw puzzle aficionado. But this is not just a complicated mental game, befitting the modern human brain. Through refitting and experimental replication of flint-knapping we can come as close as we can to seeing our predecessors’ thought processes in action.
A refitted core from Maastricht-Belvédère, Netherlands. The flakes removed from a single Levallois core have been painstakingly fitted back together.
A Levallois blade from Morfi, Greece. The left edge retains a thin sliver of the rough external surface of the lump of stone from which the tool was knapped. The knapper did that intentionally, as the rough surface provided a ready-made blunt edge to hold the tool. In Bordes’ typology, this is a naturally backed knife. The right edge was the sharp cutting edge (length approx. 78 mm).
The choices of the toolmaker do not start when the knapper sits down to start knapping and do not end when the end product has been successfully made. The knapper chooses a lump of raw material appropriate for what he/she wants to make on each particular occasion. Going even further back, the knapper chooses what raw materials to acquire, what sources to acquire them from, how much time and energy to invest in acquiring good raw materials, whether to do the knapping at the raw material source or at a ‘home base’ site or a hunting site. Then, before each strike, the knapper is presented with a situation to assess and with a number of alternative ways to proceed.
In some Palaeolithic sites we can reconstruct knapping ‘journeys’ in enough detail that we can see at each step what situation the knapper was faced with, what were the alternative courses of action available and what was the action chosen by the knapper.
All these choices are linked to when, where and for what activities the stone tools will be used, resharpened, reused and ultimately discarded. The different stages in the ‘life’ of a stone tool (raw material acquisition, knapping, use and discard) are interlinked, and at each of these stages our predecessors made choices that required planning, envisioning far ahead what the repercussions of these choices might be.
While his predecessors relied on the mere presence or absence of some characteristic tool types to draw conclusions, Bordes examined the entire lithic collection recovered from each site. He achieved this by dividing and subdividing Mousterian tools to the point of exhaustion. He defined fifty-nine different types of modified, retouched tools, plus four types of unmodified flakes and points produced by the prepared core technique. He then used simple statistics to understand which of these tool types was most prominent in each assemblage.
Of Bordes’s sixty-three tool types, twenty-one are different varieties of ‘side scrapers’, which are flakes or blades on which one or two long edges have continuous retouch along the whole length of the edge. Despite the name, side scrapers would also have been useful for cutting. Bordes’s near-obsessive categorizing highlights the perils of trying to
name an artifact type after its supposed purpose.
When Bordes looked at the relative abundance of side scrapers and the other main tool types in Mousterian assemblages from sites in northern and south-western France, he realized that they fell into distinct groups: the Charentian Mousterian (dominated by side scrapers), which he subdivided into the Ferrassie group (in which the Levallois was often used) and the Quina group (in which Levallois products were rare); the Denticulate Mousterian (dominated by tools with serrated edges which he called denticulates and notches); the Typical Mousterian (in which both side scrapers and denticulates and notches are common, in roughly equal measures); and the Mousterian of Acheulian Tradition (in which there are handaxes and backed knives, which are blunted to make them easier to hold, in addition to abundant side scrapers and denticulates). Subsequent research since the 1960s elsewhere in Europe and the Near East has supported Bordes’s findings that these are indeed distinct groupings within the Mousterian.
In A Tale of Two Caves Bordes interpreted these different variants of Mousterian lithic industries as ‘different ways of performing the same activities with different tool kits’. For him the routes of this variability were cultural. People were taught specific ways of doing things and continued to use them and pass them down the generations: ‘Our point of view is that during Mousterian times different cultures, with different traditions of toolmaking and tool-using, coexisted on the same territory but influenced each other very little.’
Bordes actually used the names of his groups of lithic industries to refer to the makers of these industries as distinct, almost ethnic groups (such as the ‘Typical Mousterians’, the ‘Denticulate Mousterians’). Stone tools become people: ‘Intermarriages are difficult to assert or refute, but in primitive societies, conservatism is usually very strong. If one supposes that a Mousterian of Acheulian Tradition man married a Quina woman, she might have gone on using the thick scrapers to which she was accustomed, but we doubt that her daughter would have done the same.’ There is little if any room here for change, innovation and spread of new ideas.
The strongest rebuttal to Bordes’s interpretation came from across the Atlantic. Lewis Binford, a larger-than-life, almost messianic figure credited with launching the ‘New Archaeology’ of the 1960s, argued in a paper with Sally Binford that the divisions Bordes identified were functional rather than cultural. Binford claimed that the different variants of Mousterian industries were made not by different groups of Neanderthals, but by the same groups while working on different activities. Certain tools were used together as distinct tool kits, each for specific functions and activities only. The activities for each site depended on the distribution of natural resources across the landscape. Binford did not elaborate much on the composition and uses of these tool kits. The two tool kits that he referred to in detail were scrapers and points on the one hand, and the denticulates and notches on the other. He suggested that scrapers and points were used for butchering and processing animals, while denticulates and notches were used for processing plants, wood and possibly cracking open skulls and long bones to access the brains and bone marrow.
Binford went further and suggested that the primary processing of animals was done mainly by the males of the group, while the processing of plants or bones was done by the females. In other words, Bordes’s ‘Quina woman’ was replaced by Binford’s ‘denticulate woman’ who made only denticulates and notches (not the more elaborate scrapers and points) and only did the humble work of processing plants or cracking open bones.
Binford’s argument was part of his distinct view of Neanderthal society, a view that for some time was very influential, at least within the Anglo-Saxon world of Neanderthal archaeology: Neanderthal men and women led fairly separate lives, with separate subsistence and other activities and little inter-group cooperation. Neanderthal life was expedient and day-to-day with little forward planning, mostly responding to immediate needs and availability of resources.
We now know that Binford was wrong on most points. Binford considered that Mousterian tools were made, used and discarded on the spot, but in fact some were transported far from where they were made. Binford thought the Neanderthals mostly scavenged dead animals or killed the weaker young, old or diseased animals, but in fact they did hunt animals in their prime. And Binford thought that the lithic points they made could not function as projectile tools, but studies have shown that they could.
Despite these faults, Binford’s approach had an enormous impact. According to Clive Gamble, the Lewis and Sally Binford paper on the Mousterian ‘was a game changer in more ways than one since it tried to do something with the stone tools other than describe and typologize them. Looking back it’s difficult to remember how contentious it was. However, so much flows from their paper in terms of a landscape archaeology for the Neanderthals that its influence on me was immediate and long-lasting.’
Bordes and Binford shared a goal of making Neanderthal archaeology more scientific, and both relied heavily on statistics. They remain influential for Bordes’s notion of analysing an entire lithic assemblage and for Binford’s emphasis on function. More recently, archaeologists such as Harold Dibble have built on their work by demonstrating that many of these different tool types can be products of a longer chain (for example, through frequent use and resharpening, single side scrapers can evolve into double side scrapers and then convergent side scrapers). Where Bordes and Bindford fell short was that their theories did not credit the Neanderthals with the ability to innovate. The Neanderthals were far more advanced than either of them imagined.
Neanderthal society
Using the evidence of stone tools and bones, which is almost all we have from this distant epoch, what can we reasonably say about Neanderthal social life? By applying the concept of forward planning to social units, we can posit that Neanderthals were better than their ancestors at anticipating what their friends and family might do in different situations. In other words, they were becoming increasingly sophisticated, enabling them to create stronger bonds and larger social units, ultimately improving their survival abilities.
This line of reasoning follows the social brain hypothesis, which we mentioned earlier in the chapter. The key insight of the hypothesis is that our species is a highly social one and that group size is a major part of our success. The intoxicating feeling of being part of a large group is something all of us experience, whether it is singing or cheering in unison at a sporting event or praying together in a religious setting. In daily life, humans not only crave this feeling of group belonging, but we depend on the cohesion of groups for our survival and advancement. A ‘group’ is not necessarily a collection of people that live together, but is perhaps better described as a network of relationships. A group is made up of people we recognize by sight or voice as being reliable friends or allies.
According to the social brain hypothesis, the increasing complexity of social skills needed to navigate human groups was probably the main evolutionary driver behind the increase in brain size that we see from the earliest Homo. Essentially, it was a positive feedback loop. As social groups became larger, more cognitive processing power was needed for an individual to succeed, which in turn led to even higher complexity. The challenge of complex social relations was a factor for both of Homo heidelbergensis’s main descendants – Neanderthals and archaic Homo sapiens – which may explain why both populations continued to evolve large brains, and with them larger group sizes, in parallel.
The author of the social brain hypothesis, Robin Dunbar, has related social group size to the size of the neocortex, which is the technical term for the outer layers of grey matter in the brain. Here is where size is not everything in brains, for modern humans have a skull that is dome-shaped, allowing for a larger neocortex than the flatter Neanderthal skull, despite the fact that Neanderthal brains were as large as or larger in total volume than our modern brains.
According to Dunbar, chimpanzees, probably like our common
ancestor with them, live in groups of fifty to fifty-five individuals. Modern humans, he argues, have brains that can maintain social relationships with 150 others, the so-called ‘Dunbar number’ which we can see across many aspects of human activity such as effective unit sizes in the military, sizes of companies or corporate departments, and in social networking.
Chimpanzees maintain group cohesion with grooming. But grooming is time consuming, giving a natural limit to the number of individuals with whom different group members can interact. We maintain cohesion through small talk, which is much quicker, allowing for bigger effective groups. Somewhere between our common ancestor with chimpanzees (around 6 to 8 million years ago) and the present, our ancestors developed language, enabling a steep increase in group size. Dunbar places this moment at about 500,000 years ago, which gives credit to Homo heidelbergensis for the development of complex speech. Genetic studies, which we discuss in Chapter Six, place this development at least that far back, and perhaps even earlier.
Using Dunbar’s theory, we can build up a picture of Neanderthal social life, which was probably quite similar to what our own ancestors were doing in southern Africa during this period. Neanderthals probably had social networks slightly smaller than ours, perhaps of 120 individuals, most likely structured along the lines of modern hunter-gatherers, who are based in clusters of about thirty people (four to six families) and have good relations with neighbouring clusters. We should note that some researchers think that Neanderthals lived in even smaller clusters, judging by the small sizes of some of the caves and rock shelters they inhabited. One of Dunbar’s students, Eiluned Pearce, argues that Neanderthals had larger eye sockets than modern humans, indicating that Neanderthal brains gave more space to visual processing at the expense of social cognition. Based on patterns of sexual dimorphism, Dunbar thinks that Neanderthals practised polygyny, with two women for each man. Curiously, these theories rarely seem to make it into museum displays and illustrations, which often emphasize anachronistic monogamous nuclear families.
The Neanderthals Rediscovered: How Modern Science is Rewriting Their Story Page 9