And then there was Levallois

7 minute read

Noble and Davidson (1996:200-201) have a great passage on the lack of relevance of the Levallois technique to interpreting ancient cognition. It has an attention-grabbing headline:

The Levallois technique: why it is not important

The usual story of the Levallois technique is that it shows the knapper intended to obtain a single large flake as an end-product. This is often argued to reflect an advanced cognitive spatial skill, since the process of shaping the core does not bear a similarity in form to the ultimate tool. For example, a handaxe can be obtained by simple reduction -- take off everything that is not part of the mental model of a handaxe, and what is left is your tool. A Levallois flake is at no time prefigured by the shape of its core; the method is in that sense a sequence of steps that are arbitrary with respect to any mental model of the flake itself.

Noble and Davidson don't believe the story. Their reasons can be paraphrased as follows:

  1. If the large Levallois flake is the goal of the entire process, then it is hard to explain why in some cases this flake was discarded along with the core.
  2. The basic Levallois core was present in earlier assemblages back to the Oldowan, so it is a relative common product of knapping and unlikely to represent a novel cognitive development.
  3. The reduction sequence leads to an extravagant waste of raw material, if the large Levallois flake is assumed to be the ultimate goal.
  4. "Waste" flakes from reduction of Levallois cores appear to have been taken away and used themselves in many instances, even when Levallois flakes were left behind.
  5. Unretouched Levallois flakes have a low proportion of use-wear, particularly compared to the "waste" flakes.
  6. Levallois flakes are not standardized in shape or size, making it seem less likely that they were the result of deliberate planning.

The received wisdom has one thing going for it: it implies that prepared core techniques become more common because of their manifest advantages. If they are just better, then that superiority explains why people adopt them and abandon large bifaces.

But this also is a weakness when considered in detail. People didn't abandon large bifaces when they started making Levallois flakes. There were a couple hundred thousand years of gradual technological overlap, and large bifaces recur in several later industries, including in Australia.

The assumption that there is a technological progression from Lower Paleolithic/ESA to Middle Paleolithic/MSA masks the functional qualities of the respective toolkits. That means that the question, "Why would anyone want a handaxe?" may appear to be much more mysterious than necessary.

Let's consider some options:

1. The techniques leading to large flakes have slight advantages over large biface manufacture; either because materials are used slightly more intensively, or they are slightly lighter to carry, or they function better. Over many millennia starting 300,000 years ago, people slowly wised up to the fact that they could save 20 minutes a day (or probably less) in flaking, quarrying, or butchering by using the slightly better tools.

This is pretty much the traditional account. It also makes very little sense.

2. Bifaces had an enigmatic function that ultimately disappeared. Maybe females used them as a marker for mate choice. Maybe males used them in initiation rites. Maybe they dropped them on unsuspecting animals out of trees. Maybe they chucked them discus-style into herds of watering animals. In any event, this function slowly faded away and later people focused exclusively on less romantic tools.

A little-noticed fact is that all these types of "enigmatic function" explanations are actually hypotheses about social organization. Either they assume that bifaces are social markers, or they assume that biface-making hominids had unique hunting styles that were ultimately lost. But it seems implausible in either case that the social factors would remain static over a million years. Every non-anatomical social marker we know of in modern humans and chimpanzees varies widely between local and regional populations. And organizational hunting techiques if anything are more complicated and harder to communicate and maintain than tool manufacturing methods. This would argue that organizational hypotheses are poorer explanations for stasis than simple functional hypotheses.

In any event, these kinds of hypotheses are attractive because they can be made compatible with the large biface accumulations, the fact that many were apparently discarded without having been used, and the weight and heft of large bifaces. They are less consistent with the morphological variability among the bifaces (which include cleavers, after all), their spatiotemporal distribution (including apparent differences between wooded and open habitat), and the fact of their ultimate disappearance, which remains arbitrary and unexplained under any of these hypotheses.

I am especially puzzled by one contradiction: handaxes have a pancontinental distribution (and were even made occasionally far from the otherwise contiguous geographical range of West Asia, Africa, and Europe), but they are rarely found more than a few kilometers from their raw material source. It is unlikely enough that a single cultural tradition could be maintained across three continents with frequent long-distance dispersals -- indeed, it would be unparalleled by anything older than the last few thousand years (and certainly not by any nonhuman primate species). It seems completely implausible that such a tradition could be maintained in the absence of frequent long-distance dispersals. Yet that is precisely what we appear to have in the Acheulean.

That brings us back to function. Large bifaces occurred across such a large range because they were easily independently invented time and time again. Such independent invention may have been emulation (i.e., invention of manufacturing methods with knowledge on the ultimate goal) based on the chance encounter with bifaces (manned or unmanned). Or it may have been complete functional convergence (arriving at the shape of a large biface by trial and error as a most-functional object). They were on occasion invented far from their spatiotemporal heyday, because they simply worked.

This leaves unanswered exactly how the presumably useful functions of bifaces were replaced by people who didn't make many of them.

3. Bifaces were cores. Later people used different reduction sequences, which generated different kinds of cores.

This answers the function question as a negative -- the "function" of bifaces didn't need to be replaced, because with respect to actual use, there is no functional difference between a large biface and a flake. But the role of the biface as a curated (and discarded) object was replaced by superior reduction techniques.

But the biface-as-core hypothesis recurses back to the first pair of hypotheses and their problems. Why did people abandon the biface-core reduction technique and start knapping differently? Noble and Davidson (1996) argue that other techniques were unlikely to have been more efficient in generating larger numbers of useful flakes or less raw materials, particularly if prepared core techniques were the preferred replacements.

And if there is anything that characterizes later assemblages compared to the Acheulean, it is greater variability. This increase in variability would seem to argue not for an increase in efficiency, but instead for the relaxation of some constraint.

The idea of a relaxed constraint also allows for the later persistence of large bifaces in many assemblages, since they would continue to persist where the constraint was still extant. But this would imply an interesting sort of social heterogeneity. I don't conceive of it as a cognitive constraint, partly because I don't think that cognitive heterogeneity among contemporary Middle Pleistocene hominids would have persisted over the necessary timespans, and partly because large bifaces recur in recent archaeological contexts.

But even in the absence of cognitive differences among contemporary humans, a relaxed constraint tends toward some aspect social organization as a cause.

And if we must depend on social organization anyway, then there is no real reason to deny handaxes a primary function. The handaxes-as-cores hypothesis is tempting precisely because it promises to explain their persistence and ultimate disappearance without recourse to their functional qualities. But if this hypothesis itself must reflect back onto the difficult-to-explain (and archaeologically invisible) aspects of social organization, then it is no more parsimonious than any of the functional hypotheses.

So we need a hypothesis that draws in as many disparate observations as possible. After all of this meandering thought, I think I may have one. We'll have to see if it works out.


Belfer-Cohen A, Goren-Inbar N. 1994. Cognition and communication in the Levantine Lower Paleolithic. World Archaeol 26:144-157.

Noble W, Davidson I. 1996. Human evolution, language and mind. Cambridge University Press, Cambridge UK. Amazon