1. Evolutionary Anthropology 109
ARTICLES
On Stony Ground: Lithic Technology, Human
Evolution, and the Emergence of Culture
´
ROBERT FOLEY AND MARTA MIRAZON LAHR
Culture is the central concept of anthropology. Its centrality comes from the fact human evolution different and what it
that all branches of the discipline use it, that it is in a way a shorthand for what is that it is necessary to explain. It is at
makes humans unique, and therefore defines anthropology as a separate disci- once part of our biology and the thing
pline. In recent years the major contributions to an evolutionary approach to that sets the limits on biological ap-
culture have come either from primatologists mapping the range of behaviors, proaches and explanations. Just to
among chimpanzees in particular, that can be referred to as cultural or “proto- add further confusion to the subject, it
cultural”1,2 or from evolutionary theorists who have developed models to account is also that which is universally shared
for the pattern and process of human cultural diversification and its impact on by all humans and, at the same time,
human adaptation.3–5 the word used to demarcate differ-
ences between human societies and
groups. As if this were not enough for
Theoretically and empirically, pa- that paleoanthropology can play in any hard-worked concept, it is both a
leoanthropology has played a less the development of the science of cul- trait itself and also a process. When
prominent role, but remains central to tural evolution. In particular, we want treated as a trait, culture can be con-
the problem of the evolution of cul- to consider the way in which informa- sidered to be the trait or the means by
ture. The gap between a species that tion from stone-tool technology can which that trait is acquired, transmit-
includes Shakespeare and Darwin be used to map the pattern of cultural ted, changed, and used (that is,
among its members and one in which evolution and thus throw light on the learned, taught, and socially passed
a particular type of hand-clasp plays a nature of the apparent gap that lies on). It exists in the heads of humans
major social role has to be significant. between humans and chimpanzees. and is manifested in the products of
However, that gap is an arbitrary one, First, we discuss the various meanings actions. To add one further dimen-
filled by the extinction of hominin of the culture concept and the role of sion, culture is seen by some as the
species other than Homo sapiens. Pa- paleoanthropology in its use. Second, equivalent of the gene, and hence a
leoanthropology has the potential to we look at how stone-tool technology particulate unit (the meme) that can
fill that gap, and thus provide more of can be used to map cultural evolution be added together in endless permu-
a continuum between humans and and provide insights into the cultural tations and combinations, while to
other animals. Furthermore, it pro- capacities of different hominin spe- others it is as a large and indivisible
vides the context, and hence the selec- cies. Third, we consider the inferences whole that it takes on its significance.
tive environment, in which cultural that can be made from stone-tool In other words, culture is everything
capabilities evolved, and so may pro- technology for the timing of major to anthropology, and it could be ar-
vide insights into the costs and bene- events in cultural evolution. gued that in the process it has also
fits involved in evolving cultural adap- become nothing.3,5–10
tations. The pervasive nature of the culture
In this paper we focus on the role EVOLUTION, CULTURE AND concept means that evolutionary an-
ANTHROPOLOGY thropology must also tackle the prob-
lems it throws up. This is not the place
Culture in Anthropology either to argue that the concept
Robert Foley and Marta Mirazon Lahr are
´ Culture is the jam in the sandwich should be abandoned as of little or no
at the Leverhulme Centre for Human Evo- of anthropology. It is all-pervasive. It analytical utility (one of us attempted
lutionary Studies at the University of Cam- is used to distinguish humans from this several years ago, to no noticeable
bridge, Downing Street, Cambridge. E-mail:
raf10@cam.ac.uk and mbml1@cam.ac.uk apes (“everything that man does that effect11) nor to come up with a cut-
the monkeys do not” (Lord Raglan)) ting-edge redefinition that will clear
and to characterize evolutionarily de- away a century of obfuscation (we
Evolutionary Anthropology 12:109 –122 (2003) rived behaviors in both living apes leave that in the capable hands of the
DOI 10.1002/evan.10108
Published online in Wiley InterScience and humans. It is often both the ex- Editor of Evolutionary Anthropol-
(www.interscience.wiley.com). planation of what it is that has made ogy). Rather, we wish to consider how

2. 110 Evolutionary Anthropology ARTICLES
TABLE 1. Cultural Capacity Is Cognitively Based, but Is Correlated With a Number of Manifestations in the Realms of
Learning, Social Organization, Symbolic Expression, and Patterns of Tradition. These Expressions May in Turn Be Visible
in the Archeological and Fossil Record
Broad Correlative Components of Culture Potential Paleobiological Manifestations
Learning capacity Technology and technological variation
Brain size?
Social organization and structure Archeological density, structure and distribution
Sexual dimorphism in fossil hominins
Nonecologically functional elements of material culture
Traits associated with symbolic thought Brain size?
Anatomical basis for language
Variation in material culture
Tradition maintenance and change Regional variation and longevity of archeological components
those aspects of anthropology that fore be a diachronic process. Compar- sistence through time). The possible
deal with the deep past of the human isons between two living species, hu- paleobiological correlates of these are
lineage—paleoanthropology—might mans and chimpanzees, can only shown in Table 1.12
throw light on the evolution of culture examine outcomes, not the actual pro-
and the role it may have played in cess of transition. This must be in- EVOLUTIONARY HISTORY IN
human evolution.12 ferred. The actual development of
THE STONES: WHAT CAN THEY
The problem in attempting this is more and more culture-bearing homi-
that the sources of such evidence are nins must have occurred among spe- TELL US?
limited, especially if little recourse is cies that are now extinct, to whom our We can now turn to stone technol-
made either to analogical or phyloge- only access is through the fossil and ogy. From over two million years
netic inferences from chimpanzees archeological record. lithic artifacts provide a rich and du-
and other primates or extrapolation To search for something in the fos- rable source of information about the
back from ethnography and psychol- sil and archeological records requires behavior of extinct hominins, and
ogy. Paleoanthropology is limited to knowing what one is looking for, so thus greatly expand on the anatomical
the archeological record for the evi- that a consideration of definitions of fossil evidence. The question to ask is
dence it throws either on hominin culture cannot be entirely avoided. what sort of information can be de-
cognition, and hence culture, or else Definitions of culture largely fall into rived from stone tools?
on the cultural manifestations of be- two broad groups. Either they involve Archeologists have basically come
havior. In practice, this means using the actual end products of behaviors up with two answers to this question.
the record of stone tools, the primary that are inherently human (technol- On one hand, patterns in technology
source of information about the be- ogy, for example) or they focus on the have been used to reconstruct popula-
havior of prehuman hominins. processes that produce these out- tion histories, in a sense to construct
comes—that is, the cognitive under- phylogenies of species and popula-
pinnings. Most recent approaches tions (cultures, in other words). Stone
Culture and
have concentrated on the latter or, in tools were, in effect, treated as popu-
Paleoanthropology other words, trying to get into the lation markers. This may be consid-
There are two reasons why both minds of extinct species and popula- ered the phylogenetic and historical
evolution and paleoanthropology are tions. This can only be done in terms approach. On the other hand, stone
central to any discussion of culture. of correlates. Most definitions of cul- tools can be and have been interpreted
The first is that the distinction be- ture involve three core elements: those as adaptive markers, often with little
tween humans and other species is associated with learning, its depth or no phylogenetic signal, because
usually drawn in some way around and extent, or the ability to acquire they are endlessly thrown up conver-
the concept of culture—put simply, new information independent of a gently by the demands of the environ-
we have it and they do not. Chimpan- tightly constrained genetic basis; ment and social organization, which
zees chipping away at the margins of those associated with social organiza- thus reflect variability in behavioral
tool making or grappling with the ru- tion and complexity; and those asso- response. This can be termed the
diments of American sign language do ciated with symbolic thought, both its adaptive function approach.
not really change this state of affairs. underlying cognitive basis and its By and large, these two approaches
Given the fact that humans must have communication. In addition to these have been seen as alternatives, and to
evolved from an acultural organism to core elements is the extent to which be in conflict with one another. Fur-
one that possesses such capacities the behaviors derived from these ca- thermore, from a historical perspec-
means that the evolution of culture is pacities are either capable of change tive, the adaptive function approach
a major challenge to evolutionary the- and variability (a characteristic of cul- has generally supplanted and suc-
ory. The second, related, aspect is that tural systems) and have a means of ceeded the phylogenetic and historical
the evolution of culture must there- being maintained as traditions (per- approach, and has become the con-

3. ARTICLES Evolutionary Anthropology 111
sensus on which most Palaeolithic ar- stone tools were still seen as markers ings of the populations who made
cheology operates. However, it is worth of peoples as they ebbed and flowed them, reflects the demands of the en-
considering briefly the strengths and across the Palaeolithic landscape. The vironment and the responses of the
weaknesses of each approach. high tide of the phylogenetic and his- populations to those demands within
torical approach occurred when it was the constraints of raw-material avail-
possible to draw simple boundaries ability.
Phylogeny and History: around typological and technological The switch in emphasis was encap-
Human Evolutionary History clusters and to associate them with sulated in the Mousterian debate of
From Stone Tools cultural history and narrative. Thus, the 1970s, when Binford argued that
the cultures of the Upper Palaeolithic, the variation in the frequencies of tool
The idea that human evolutionary for example, were essentially analo- types in the rock shelters of the Dor-
history might be reflected in stone- gous to ethnographic units, an anal- dogne and the Levant reflected differ-
tool typology is one of the oldest in the ogy that was sometimes drawn all too ent activities being carried out, rather
discipline and, in one form or an- explicitly.14 than the movements of different peo-
other, has been a persistent theme This “from technology to culture to ple. The form of stone tools and their
over the last one hundred and fifty people to history” approach has been frequencies in assemblages have been
years or more. When Frere recognized subject to many criticisms, and is seen increasingly as the result of envi-
the stone tools discovered in the eigh- largely associated with work by arche- ronmental and ecological demands
teenth century as the product of hu- ologists in the first half of the twenti- and opportunities. Concomitant with
mans, and at the same time recog- this view is the corollary that if the
nized that they were very “primitive,” signal in the shapes of stone is func-
he was drawing the first of many such
conclusions. Stone-tool typology
The idea of stone tools tion, it could not at the same time be
phylogenetic and historical.
could be seen to reflect the stages of as the markers of To this strongly ecological ap-
human history, from the first simple chronology gradually proach has been added an additional
flakes and cores through to the So- element, that of the constraints of
lutrean points. During the first part of fell into disrepute, stone as a raw material and the pro-
the twentieth century, this became especially as it was cess of knapping itself. It is clear that
formalized in the schemes of Breuil,
Burkitt, and Bordes.13 recognized that globally in some parts of the world good lithic
materials are abundant, and in others
The phylogenetic and historical ap- it was hard to maintain scarce. The strategies of stone tool
proach generally encapsulated two
basic components. The first was that if the model of universal manufacture would therefore be ex-
pected to reflect this. The classic ex-
stone tools were similar, then they stages and that there ample of this view has been the in-
were made by the same sort of people,
usually taken to mean people belong-
was not necessarily any creasingly popular interpretation of
the Movius Line as a raw material
ing to the same culture, with greater chronological boundary within the Old World.17,18
or lesser implications for ethnic
groups, depending on the time scale
consistency to the The way in which stone tools are
made—through a process of core and
involved. The second was that the pattern of change. flake reduction—is also important. It
level of sophistication or complexity has been argued that the differences
of the tools reflected the cognitive or among typological elements are the
cultural status of the population con- product of different degrees of reduc-
cerned, usually more or less advanced tion, and that, for example, a few
within the framework of the time. eth century. The move to a greater
emphasis on adaptation and, more re- more blows and one type is trans-
When these two components are put formed into another. Tool-type fre-
together, one has a model for explain- cently, raw-material constraints, has
greatly altered the way Palaeolithic ar- quency thus reflects use and the need
ing prehistoric change in terms of the to retouch more or less. From an evo-
movements of peoples through their cheology has been done and how the
past is interpreted. lutionary perspective, the adaptive
particular set of tools with a process function approach sees homoplasies
of evolution toward greater cultural (convergent evolution brought about
and, by implication, cognitive com- Adaptation and Function: through a combination of selection
plexity. and constraints) as being rife, and
The idea of stone tools as the mark-
Information From Design
therefore the phylogenetic signal of
ers of chronology gradually fell into The alternative to the idea that stone tools as being very low.
disrepute, especially as it was recog- stone tools reflect population and thus
nized that globally it was hard to evolutionary history is that of adap-
maintain the model of universal tive function, and is the consensus
Back to Population History
stages and that there was not neces- view of archeologists today.15,16 Vari- In recent years, however, there has
sarily any chronological consistency ability in stone tools, rather than re- been a resurgence of interest in the
to the pattern of change. Nonetheless, flecting the social and cultural group- interpretation of archeological mate-

4. 112 Evolutionary Anthropology ARTICLES
rials in an evolutionary, in the sense of be expected to reside in the character- ideas we will develop here are one at-
phylogenetic, perspective.19 This can istics of the various species and not to tempt at disentangling these signals.
be seen in areas of direct interest to exhibit a great deal of sensitivity in re- First we look for the presence or ab-
human evolution. One example is the lation to the environment. This allows sence of a correlation between biolog-
association of the Aurignacian indus- us to consider whether stone-tool tech- ical evolution, based on morphologi-
tries with the dispersals of modern nology covaries with phylogeny and cal affinities, and technological change,
humans into Europe and, conversely, taxonomic status or with the environ- based on the distribution of technolog-
the issue of whether there is a link ment, and thus provides an empirical ical modes. Second, we use this derived
between Neanderthal populations and route into the problem. relationship to consider whether there
the Mousterian in general and the In summary, therefore, we need to is an association between cultural out-
Chatelperronean in particular.20,21 A consider stone tools both in an envi- put and the species involved, and where
further example is the suggestion by ronmental context and in the context technological change occurs in relation
Klein22 that the dispersal of Homo hei- of phylogeny. Both history and ecol- to biological change. Finally, we con-
delbergensis into Europe is associated sider how these might relate to inferred
with the Acheulean. We have also pro- cognition. Central to our argument is
posed that stone tools are markers of that while environment is shaping the
hominin geographical patterns,23 The extent to which technological demands, the nature of
both in the long-term persistence of hominins’ behavioral response is cir-
the Movius line and in the spread of
hominins might have cumscribed by their cognitive abilities.
Mode 3 or prepared core technologies possessed a greater or Thus, the link between technology and
in Africa and Europe as part of a dis-
persal of later archaic populations, as
lesser degree of cultural phylogeny is crucial for determining
the pattern of cultural evolution.
well as modern humans.24 capacity might be
The Evolution of Culture
expected to be The Pattern of Hominin
Evolution
Through Stone Tools: Which reflected in the extent to
To provide a framework, we can
Approach? which we can see a briefly outline the pattern of hominin
Given these two contrasting ap- good fit between the evolution from the origins of the ge-
proaches to the information poten-
tially locked in the stone tools, we ask
environment and nus Homo. Figure 1 shows the distri-
bution of proposed genus Homo taxa
which one can give the most useful technology. Here the by time and geography. The earliest
insights into the problem of the evo- proxy for culture is thus Homo, as well as the australo-
lution of culture, and thus make use of pithecines, are excluded: Although
the archeological record within the taken to be those there is clear evidence that they did
field of anthropology more generally. aspects of the various make stone tools,25–27 this primarily
Perhaps the common-sense answer is suggests either that Mode 1 technolo-
the adaptive functional approach. definitions that gies are plesiomorphies of Homo, be-
This would certainly be the preferred emphasize the ing developed among one or more
option for most archeologists, as it australopithecine lineages, or else an
represents the prevailing paradigm behavioral apomorphy at the base of Homo. The
for the analysis of stone tool variation. manifestations of subsequent distribution of Mode 1
More importantly, as culture is pre- technologies shows the diversification
sumably an adaptation, then it is only
culture, variability, and and geographical radiation of the de-
natural to use an adaptive approach to a high rate of change. scendants of Homo ergaster or possi-
identify it in the past. The extent to bly earlier members of Homo. Among
which hominins might have possessed these geographically widespread
a greater or lesser degree of cultural members of Homo there appears to be
capacity might be expected to be re- considerable diversity, with a distinc-
flected in the extent to which we can ogy are important, as is the case in tive pattern to be found in Eastern
see a good fit between the environ- most evolutionary problems. Testing Asia that has led some authorities to
ment and technology. Here the proxy the various possibilities requires a du- distinguish between an African lin-
for culture is thus taken to be those alistic approach. eage (H. ergaster) and an Asian one (H.
aspects of the various definitions that erectus).28,29
emphasize the behavioral manifesta- STONE TOOL TECHNOLOGY Newer finds, such as those from
tions of culture, variability, and a high Dmanisi,30 Ceprano,31,32 and Buia33
rate of change.
AND HUMAN EVOLUTION support this perspective, although
If, on the other hand, culture is seen Against this historical background, others such as the material from
as a cognitive state reflecting the ability we propose that embedded in the Baka34 have been employed to ques-
of the mind to generate new behaviors, Palaeolithic record are the signals of tion such a distinction. The evolution-
then this can be something that might both adaptation and phylogeny. The ary changes that occur from a little

5. ARTICLES Evolutionary Anthropology 113
size different traits, including means
of flake production, typological forms
and frequencies, metrical variation,
core reduction sequences, and mi-
crowear patterns. The geographical
and chronological scale of variation in
each of these is very variable, and
many show high levels of local, small-
scale diversity rather than the large-
scale one that we associate with homi-
nin phylogeny biologically. We argue
that in terms of mapping the general
patterns of change and stability, what
is needed is a scheme that operates on
a global scale and reflects broad-scale
change rather than local site varia-
tion. To make a biological compari-
son, we need a system that has high
interpopulation variation relative to
intrapopulation variation. Against
this criterion, the most appropriate
classification system is that of techno-
logical modes, the major forms of
lithic production (see Box 1).
The principles have been developed
elsewhere,23,24,39 but in brief consist
in recognizing general technological
traits and treating them cladistically.
These traits refer to the basic means
by which the stone tools were made
and the broad nature of the artifactual
outputs. Using Clark’s modes,40 five
basic technologies have been recog-
nized: Mode 1 being chopping tool
and flake industries; Mode 2 being the
production of bifaces and bifacially
worked handaxes; Mode 3 being pre-
Figure 1. Chronological and geographical distribution of recognized taxa of Homo.
pared core technology; Mode 4 being
lamellar or blade technology; and Mode
5 being microliths. Although there are
more than 0.6 Myr have led to the present in Africa probably from continuities between them, they ex-
view that there is a new taxon, H. hei- 150,000 years ago, but occur, presum- press more complex ways of making
delbergensis, which had a larger cra- ably through population expansions, stone tools, leading toward greater con-
nial vault and a generally more mod- in other parts of the world consider- trol and a more effective use of raw
ern appearance, although retaining ably later: 100,000 years ago in West- material to produce particular end
the extreme robusticity of the Lower ern Asia, 60,000 years ago in Austra- products. They are particularly suitable
Pleistocene Homo species.35 This lia, and around 40,000 years ago in to be considered cladistically and so
taxon is found in Africa and Europe, Mediterranean Europe and Eurasia.38 phylogenetically, because they are built
and to some it may also be present in upon each other, and incorporate some
East Asia. A further element of diver- Technological Modes, of the elements of “descent” that are
sity can be added to this essentially Hominin Phylogeny, and the essential to an evolutionary approach.
Middle Pleistocene pattern with H. It is important to emphasize that one
Scale of Environmental
antecessor, known from Spain.36 Fi- of the reasons that the technological
nally, the terminal Middle Pleistocene Variation modes are appropriate for evolution-
and the earlier parts of the Upper How do stone tools map on to a ary analysis is that they stress the
Pleistocene show the evolution of two phylogeny of the genus Homo? This most derived elements in an assem-
highly encephalized and derived raises the question of how we “mea- blage, for it is well-known that even
forms of hominin, Neanderthals in sure” technological diversity. There is after the development of more derived
Eurasia and modern humans in Af- no generally agreed means of doing modes, more “primitive” ones, in the
rica.37 The latter, H. sapiens, are this, as different approaches empha- cladistic sense, persist.

6. 114 Evolutionary Anthropology ARTICLES
Box 1. Clark’s Technological Modes
Clark, working in the context of a to be relatively small compared to the variants. The Acheulean is known
plethora of archeological cultures and size of the cores, and to lack, both on from Africa from dates close to 1.5
terminological diversity, attempted to the cores and the flakes, significantly million years ago, although it often is
provide an overarching framework for invasive retouch. This mode resulted difficult to draw a line between this
summarizing variation in Paleolithic in relatively little diversity of tool and the developed Oldowan Mode 1
and Mesolithic lithics on a global forms, relatively little by way of core industries. The bulk of well-docu-
scale. He suggested that across the reduction, and lack of any prepara- mented African Acheulean sites are
range of lithic assemblages there tion of the striking platforms. Mode 1 less than 1 million years old, and usu-
could be seen some generalities that occurs extensively throughout the ally belong to the Middle Pleistocene.
related to the way in which the stone Old World over much of the Pleisto- In Europe, Western Asia, and the In-
tools were actually manufactured— cene and well into the Pliocene in dian subcontinent, the dated
hence, the term modes. Clark had in sub-Saharan Africa. The African Acheulean sites fall mostly into the
mind the technological modes of pro- Mode 1 industries are primarily Plio- Middle Pleistocene, although there
duction for the Stone Age, upon cene and Lower Pleistocene, may be some evidence (at Ubeidiya)
which were superimposed the variet- whereas in Eastern Asia they persist that it sporadically occurred earlier.
ies brought about by cultural prefer- until the Upper Pleistocene. They also There has been prolonged contro-
ence, economic need, and raw mate- occur in the Middle Pleistocene in Eu- versy over the presence of true Mode
rial availability. His modes provide a rope. 2 industries in Eastern Asia. Although
basic framework for grouping and Mode 2 saw the development of there is some evidence for bifacial
separating stone-tool assemblages at two elements, although of course it stone tools in that region, there is
a general rather than specific level. would have been possible for these to nothing truly like the recurrent
They are described here in outline occur independently. The first of Acheulean of the west. It is this dis-
form, with their broad geographical these was the ability to strike off rel- tinction that is represented by the
and temporal distribution. atively large flakes so that they would Movius Line.
Clark’s modes were based essen- have some of the size properties of Mode 3 represents a major shift in
tially on the way in which the basic cores, but with a narrower cross-sec- the output of lithic production, al-
flake-core relationship occurred. tional area, and thus be suitable for a though it shares with Mode 2 ele-
Mode 1, comprising the Oldowan and greater amount of invasive retouch. It ments of the way tools are produced.
Asian Pebble Tool and Chopping Tool was this that constituted the second The key difference is that the core is
Traditions, constituted the simplest development, for it became possible prepared prior to striking off a major
mode of production, the striking of a to retouch the resulting flakes in such flake as a means of having greater
flake off a core. The number of flakes a way that secondary flakes were re- control over the shape and thickness
could vary, but what held this system moved across the whole surface of of the flake. The actual means of
of production together was the sim- the flake and on both sides. The result preparation, however, is probably
ple platforms and lack of preparation was the bifacial tradition that is rep- similar to that used in the production
involved. The flakes struck off tended resented by the Acheulean and its of handaxes. The outcome is a much

7. ARTICLES Evolutionary Anthropology 115
Box 1. Clark’s Technological Modes (continued)
more diverse set of finished tools, elongated blades with narrow cross- flakes and blades that are retouched
and hence a greater potential for vari- sections, which then are reworked and worked into various shapes in
ability and a greater emphasis on extensively into diverse sets of sub- some contexts or are used as com-
smaller items. Mode 3 constitutes the sidiary tool types. Although elongated posite unmodified tools in others. Mi-
technologies of the European Middle flakes (that is, blades) are produced croliths are widely known in the later
Palaeolithic and the African and In- by the Mode 3 technologies, the parts of prehistory. They form the ba-
dian Middle Stone Age. Its presence Mode 4 system is different in that it is sis of the African Later Stone Age
in the Middle Pleistocene of Eastern based on prismatic cores. Conven- from approximately 30,000 years
Asia is disputed, but it may have had tionally, Mode 4 industries are asso- ago. However, there may have been
a more extensive eastern distribution ciated with the Eurasian and North earlier occurrences of this mode (for
in the Upper Pleistocene. African Upper Palaeolithic and occur example, the Howieson’s Poort in
Mode 4 continues the trajectory of late (after 50,000 years) in the Upper southern Africa around 80,000 years
Mode 3 in the sense that it is con- Pleistocene. Blades are also known ago). Microliths are also known in
cerned with producing pieces off a to occur in earlier deposits, for exam- Southern Asia from around 30,000
core with the shape of those pieces ple in the Kapturin Beds in Kenya, years ago, more widely across Eu-
being determined by the way in which and in the early Upper Pleistocene of rope and Asia in the latest parts of the
the core has been prepared. In this Western Asia and Northern Africa, but Pleistocene, and in the early Holo-
case, the preparation is designed to these are seldom prismatic. cene (the Mesolithic). Mode 5 indus-
produce long flakes and results in cy- Mode 5 involves microlithic tech- tries are also known in the mid-Holo-
lindrical prismatic cores and fine, nologies: the production of very small cene in Australia.
Figure 2 shows the distribution of two trees is remarkably similar: Both rather than local scale of variation.
the technological modes represented show deep African/Asian clades and These are, of course, two well-known
in phylogenetic terms. It is perhaps relatively prolonged longevity of lin- and established facts, and it would
striking that the overall shape of the eages. This confirms the continental perhaps be surprising if there was no

8. 116 Evolutionary Anthropology ARTICLES
gence or at least a period of contact
and cultural diffusion around 300 Ka.
Elsewhere we have proposed that Ne-
anderthals and modern humans may
have shared a more recent Middle
Pleistocene ancestor than H. heidel-
bergensis, a population we named H.
helmei.24 It should be noted that our
use of H. helmei differs from that made
later by McBrearty and Brooks48 to re-
fer to the immediate African ancestor of
modern humans only.
The preceding evidence suggests
that there is a strong but not entirely
straightforward relationship between
phylogeny and technological modes.
This may seem to indicate that in
terms of the two approaches dis-
cussed earlier, the phylogenetic and
historical approach is the most con-
sistent with these data. This may sug-
gest that there is not a strongly adap-
tive element to technology. This is
misleading in two ways. The first is
that while the technology is adap-
tive—that is, carrying out particular
functions that enhance survivor-
ship—it is strongly mediated by the
cognitive capacities of those homi-
nins, who appear to have been limited
at least in terms of their ability to in-
novate and vary their productions.
This enhances the idea that the stone
tools are providing insights into the
evolution of the cognitive basis for
culture. The second way in which we
may be misled is if the approach
Figure 2. Chronological and geographical distribution of lithic technologies in terms of through modes is insensitive to the
modes. scale of variation that is significant at
an adaptive level. This has been one of
the criticisms leveled at the approach
concordance given that they are sup- ones in the same region did—a mea- and can be discussed in terms of “pri-
posedly the records of the same pop- sure of heritability, as it were, among vate histories.”
ulations. How, though, does this pat- (admittedly nonreproducing) arti-
tern relate to the expected scale of facts. There is a fidelity of form that Private Histories
variation? The answer to this question defies the scale of ecological variation There are many caveats to the broad
is that the scale observed seems to and seems to suggest that the varia- interpretation of the archeological
reflect long-term phylogenetic pat- tion in stone tools as refuted in Clark’s record presented, of which the most
terns more than fine-grained adaptive modes says more about the character- important one is that the modes
ones. If environment was driving istics of their makers than the envi- clearly reflect only a small part of the
Lower Paleolithic variability, one ronments in which they were living. variability in stone tools, and it could
would perhaps expect a far more frag- There are, however, differences be- be argued that they are the only ones
mented distribution, with, for exam- tween the two. Where most interpre- that reflect this scale of variation. Ty-
ple, frequent oscillations between tations of the fossil phylogeny suggest pology, assemblage structure, and mi-
Mode 1 and Mode 2 industries as hab- a divergence between European and crowear analysis might well display
itats changed and as the availability of African lineages dating back to the diversity at either more general scales
raw material varied from region to re- middle or early part of the Middle or more local ones. This in itself
gion. This is not what is seen. Instead, Pleistocene, the shared technology of would not be surprising or necessarily
the best predictor of what an artifact the Neanderthals and modern hu- a problem with this evolutionary his-
is going to look like is what the earlier mans (Mode 3) suggests a later diver- tory model. Microwear,41 for example,

9. ARTICLES Evolutionary Anthropology 117
might well be expected to map onto a THE EVOLUTION OF CULTURE evidence for the origins of Homo. In
very large scale of variability, as it is THROUGH HOMININ contrast, the earliest evidence for
probably the case that different stone Homo ergaster does not relate to any
EVOLUTION
tools were used for the same purposes significant change in technology, but
by different populations. In other Technology and Evolution: rather technological change occurs
words, there is only one way to skin a Correlation and Causality considerably later, when Mode 2 ap-
dead cat, but many tools that can be pears, after 1.4 Myr. It is also the case
We can put this notion into practice
used to do it. At the other extreme, the that Homo heidelbergensis, which is
by considering the relationship be-
detailed typological shape of the end- known from about 600,000 years ago,
tween the major changes in modes
product artifacts may well be ex- is not associated with a new techno-
and the appearance of new taxa as
pected to display local variation, as logical mode, although there is some
shown in the fossil record.23,24,39 In
these will be influenced both by the evidence to suggest that at this time
Figure 3 a phylogeny for Homo is
availability of raw materials and there is an intensification of biface
shown, with the appearance and dis-
small-scale cultural tradition, the Pa- production. Finally, when we look at
leolithic equivalent of the different the later parts of human evolution,
ways of hand-clasping or nest-build- there is some tentative reason for sug-
ing found among chimpanzees. A pre- . . . although the modes gesting that the emergence of Mode 3
historic human example would be the
differences in detailed bone harpoon
do not tell the whole technologies in Africa may be associ-
ated with a new morphology—what
shape found among the epi-Paleo- story, they do tell an we have referred to elsewhere as
lithic populations of northern Europe, important one. This Homo helmei. However, both H. sapi-
which shared the same basic stone- ens and H. neanderthalensis make
tool technology, and which Clark used might perhaps be a their appearance in the context of
to identify social territories.42 pointer to the way in Mode 3 technologies, with Mode 4/5
We argue that although the modes only occurring tens of thousands of
do not tell the whole story, they do tell which we think about years after the first anatomical evi-
an important one. This might perhaps integrative approaches dence for modernity.
be a pointer to the way in which we To many, the complexities of the
think about integrative approaches to
to human evolution. relationship between hominin lin-
human evolution. There are many There are many sources eages and technology might lead to
sources of information about the evo- of information about the the view that there is no relationship
lutionary past, from fossils to archeol- at all. Certainly there is no simple
ogy to genetics. While ultimately each evolutionary past, from causal relationship between the devel-
must be the product of a single series fossils to archeology to opment of new technologies and spe-
of historical events, nonetheless each ciation. There is not even a consistent
may have to some extent a private his- genetics. While relationship, in the sense that techno-
tory. Genes may record events that are ultimately each must be logical change always precedes ana-
completely invisible archeologically— tomical change or vice versa. There is,
indeed, one would expect them to—
the product of a single nonetheless, an important pattern
while the stone tools might be highly series of historical that requires explanation. What is
sensitive to changes that are not seen
in cranial morphology. Indeed, as the
events, nonetheless likely is that different elements are re-
lated to different events. Speciation
number of genetic systems studied in- each may have to some or, more prosaically, the date of first
creases, it is becoming clear that while extent a private history. appearances, is a demographic pro-
they tell the same basic story, each cess, usually arising from the occur-
one does have a private history: the Y rence of small isolated populations. It
chromosome compared to mtDNA, is not inherent in this process that
beta-globin compared to Alu inser- there should be a technological or be-
tions, and so on. Different elements of appearance of the technological havioral or adaptive change. Rather,
stone-tool technology may well also modes superimposed. It can be seen this process relates to genetic diver-
have their own private histories, and that the relationship is far from gence, either through drift or selec-
these histories may be regionally and straightforward (Fig. 4). It may be tion. The major behavioral changes
chronologically specific. For this rea- that this complexity is at least partly that might be associated with any new
son, technology may well not provide due to imprecise dating, but it may species could arise on either side of
a single line of evidence and informa- also reflect to some extent the fact that that geographical boundary. Major
tion, but separate ones relating to dif- while both the stones and the fossils adaptive changes, in other words, are
ferent evolutionary events—some to tell the same story they are sensitive to not necessarily related to speciation.
speciation, some to dispersals, some different parts of it. For example, we What they may be associated with are
to behavioral grade shifts, some to can see that the emergence of stone- dispersals. That is, where technology
cognition, some to ecology. tool technology predates the current confers a major adaptive advantage it

10. 118 Evolutionary Anthropology ARTICLES
the Mode 2 industries, even subdi-
vided into flake-based and nodule-
based, are characteristic of particular
periods and continents, and that they
do not change much. It is perhaps a
forgotten wonder of the archeological
world that a French-trained archeolo-
gist who knows of nothing but the
Dordogne could go to the Cape of
Good Hope and recognize the arti-
facts and mode of production. What
does this tell us about culture? Two
things come to mind. The first is that
across time there is clearly an increase
in the complexity of the means by
which tools are made, involving both
more careful material selection, more
forethought in the approach to pro-
duction, and the potential for a
greater diversity of outcomes. Unfash-
ionable as it may be, this can be de-
scribed as a progressive trend. How-
ever, the question to ask is where
across this trend are significant
changes occurring. This is not just
“chronological” variation. The modes
persist much longer in some places
than others (for example in Eastern
Asia with Mode 1), suggesting that the
evolution of the underlying cognitive
capacities of the hominins was not
uniform across the world. If the
modes reflect culture or cultural ca-
pacity, then culture is not evolving
uniformly across the world’s hominin
population. At present there is insuf-
Figure 3. Comparison of chronological and geographical distribution of lithic modes and ficient data across Asia to understand
Homo taxa. the details of this and whether it is a
case of isolation or local selection, but
as a problem it emphasizes the need
leads to a geographical range expan- are deeply stable. Despite minor typo- to situate the archeological record on
sion, and this will be visible: Hence logical variation and raw-material the hominin phylogeny. It is no longer
the often apparently rapid widespread constraints, there is little doubt that possible to refer to generalized evolu-
distributions of novel technologies.43
This may explain why the appearance
of modern humans in Europe is asso-
ciated with a new technology, the Au-
rignacian or Upper Paleolithic, but the
anatomical features associated with
these populations have been present in
Africa for as much as 150,000 years.44
The Evolution of Culture:
Inferences From Technology
What, though, can we learn about
cultural evolution from modes? The Figure 4. Relationship between technological change and lineage change among hominins.
most obvious point is that these tech- The left diagram shows the major lineages of Homo and where Mode changes occur within
nological systems of fossil hominins them, while the right diagram shows how mode changes relate to “species” changes.

11. ARTICLES Evolutionary Anthropology 119
tion of cultural capacities within the cesses of demographic expansion into Between these two extremes lie
genus Homo. various environments, and probably Modes 2 and 3. The development of
The second cultural aspect is the reflect the processes described in Mode 2 at one level seems to show a
stability of the modes across the Pleis- Shennan’s density model of cultural major change: the ability to strike off
tocene, which has been extensively explosions.49 large flakes and invasively retouch
discussed here and elsewhere. In one At the other end of the technological them in a controlled way, with a per-
sense this stability mirrors a condition spectrum, the development of Mode 1 ception of the importance of final
of culture—faithful replication of sys- technologies has been seen as a signif- shape.16,50 This shift occurs during
tems— but it does so on a scale that is icant cultural evolutionary event, dis- the span of Homo ergaster. However, it
manifestly very different from that of tinguishing more advanced hominins is worth noting several points about
modern technologies. This argues ei- from apes.16 Although there is some the development of Mode 2. It does
ther for a remarkable cultural tem- experimental evidence that chimpan- not appear with H. ergaster (1.8 Myr),
plate beyond the capacities of modern zees are capable of stone fracture but several hundred thousand years
humans or for the absence of another techniques, these appear to be later; while the end product (the
cultural trait, the ability to innovate achieved with difficulty. Homo habi- Acheulean) is distinctive, it does
and make modifications. This latter lis, or whichever Pliocene hominin merge more gradually with the Devel-
possibility seems the more likely, with oped Oldowan (Mode 1); and there is
a sense that one part of the cultural a considerable contrast between the
program, imitation, was far more . . . there is some earlier forms and the later modern de-
dominant in earlier hominins than it
is in modern humans. As Byrne45 has
experimental evidence rived Mode 2 that is associated with
H. heidelbergensis, where there ap-
shown for gorillas, imitation is quite a that chimpanzees are pears to be a much greater emphasis
complex cognitive process, so this capable of stone on symmetry and regular form, espe-
does not mean that these creatures cially once access was gained to the
were not considerably more intelli- fracture techniques, flint sources of Europe. From the per-
gent and culturally competent than these appear to be spective of cultural evolution, Mode 2
living apes. does represent a major cognitive shift,
Finally, with regard to modes, we achieved with difficulty. but its full impact is a gradual pro-
can ask whether the points at which Homo habilis, or cess, not a sudden punctuated event
they change are significant events in followed by prolonged equilibrium.
the evolution of culture or are what
whichever Pliocene This seems to suggest that although
has been referred to earlier as private hominin first made stone the rate of change is glacial in com-
histories acting independently of the tools, was clearly able parison to modern cultural change, it
rest of the hominin evolutionary does show a pattern of development
record. The mode change that has at- to replicate the process that can be interpreted in terms of the
tracted the most attention recently is consistently. This does refinement of a practice.
that of Mode 4, the blade technology Mode 3 represents a different situa-
associated with the Upper Palaeo- probably represent a tion. It can be cogently argued that the
lithic.22,44,46 This has been strongly as- significant change in the basic technique of Mode 3, the prepa-
sociated with the appearance of mod- ration of the core prior to flaking, is
ern human behavior and the Out-of-
process of cultural inherent in the Mode 2 technologies,
Africa model of recent human evolution. and a “Levallois component” has long
evolution. However, as various au- been recognized as a part of many
thors have pointed out,24,47,48 it is dif- Acheulean assemblages. This has led
ficult to pinpoint a direct cognitive some to suggest that the distinction
change with this Mode. First, it is too between the two is insignificant. How-
regionally specific, essentially being first made stone tools, was clearly able ever, although the actual technologi-
confined to Eurasia. Second, it is too to replicate the process consistently. cal aspects of change may be contin-
late, having occurred well after the This does probably represent a signif- uous, the outcomes are radically
first appearance of modern humans icant change in the process of cultural different. Rather than the repetitive
and after the diversification of the hu- evolution. Strout and coworkers have and monomorphic production of
man population. If it was a cognitively used PET scans of people carrying out handaxes, instead there is the diver-
and biologically based cultural shift, stone knapping to explore the cogni- sity of flake forms. The shift repre-
then it occurred only after the major tive processes involved and shown sents a major change in the way stone
populations of the world had sepa- that these do share similarities with cores (even if the cores are large
rated, and therefore could not be a cognitive responses to tasks of a cul- flakes) are used and developed: They
universal trait of humanity. Mode 4 tural nature in the extent to which become the template from which di-
and, we argue, Mode 5 as well, are they coordinate motor control with versity can be produced rather than
important, not as markers of major other aspects of cognition, especially the end product themselves. This can
cognitive evolution, but of the pro- spatial processing. be seen in the increase in variation

12. 120 Evolutionary Anthropology
that occurs in the Middle Stone Age Modes 4 and 5. Certainly, it seems and the earlier hominins. Further-
and Middle Palaeolithic, both within that there is a strong contrast in be- more, the fact that these changes oc-
and between assemblages.51,52 With havior and apparent cognitive flexibil- cur across the time of the lineages
Mode 3 we see something that begins ity between hominins prior to H. hei- concerned suggests that this is not a
to approach the variation we would delbergensis and those after. It is case of behavioral or cognitive sta-
associate with modern cultural behav- perhaps significant that this is also the bility. The development of Mode 3
ior, and its appearance may be related period when brain-size evolution ac- (H. helmei, H. neanderthalensis, and
to other substantive changes in behav- celerated. H. sapiens) represents an even greater
ior.53 shift, with both standardization of
Who Has Culture, Whatever form and diversification of end prod-
Cultural Status of Extinct uct. Comparison across Modes 2 and
That Is? 3 suggests that there is an earlier cog-
Hominins nitive shift related to the ability to im-
We are aware that compared to the
On the basis of the preceding dis- rich tapestry of culture in the other itate and to maintain content and
cussion, we could argue that the tech- papers in this issue, our version is form (tradition?), and a later one as-
nological modes do provide useful in- somewhat stony and bare. There is no sociated with innovation. This latter
sights into the evolution of culture, web of kinship or devious monkeys, change, when viewed in the context of
but for this to be strengthened it needs no language, and no symbolic the evolution of modern humans and
to be more firmly rooted into other thought. In a way, our intent has been the amazing accretion of diversity of
aspects of human evolution. We have to trace the most basic of patterns in material culture that occurs through
shown (Figs. 1–3) that there is consid- as broad a comparative context as the last 100,000 years, suggests that
erable congruence in the broad distri- possible, so that we can see how cog- the evolution of these cultural capabil-
bution of modes and hominin popula- nitive state might map on to the radi- ities was not a single event, but cumu-
tions, but this is far from simple, and ation of hominins as seen in the fossil lative. Perhaps the most important
that not all mode transitions show the record. This has meant confining our- conclusion is one that stresses the im-
same pattern in relation to biological selves to a single source of informa- portance of looking at evolution diach-
evolution. This is summarized in Ta- tion, stone tools, and a large-scale ap- ronically: The evolution of culture is not
ble 2. proach, technological modes. Given a single step. Rather, the gap between
From this two major points emerge. this limited approach, we can see that humans and chimpanzees, between a
The first is that there is no simple re- the similarities between the fossil few termites for lunch and Beethoven,
lationship between modes and homi- record and the technological one sug- is filled with incremental steps.
nin species. For example, most tech- gest that the latter has a strong phylo- While it has been possible to gain
nology-using hominins made Mode 1, genetic signal, and that this can be insights into the cognitive states of ex-
itself an interesting insight into the interpreted as showing that the ability tinct hominins via the relationship be-
evolution of culture, suggesting a deep to generate technological solutions to tween technological modes and mor-
plesiomorphic conservatism for most adaptive problems was limited in phological affinities, it may be
of human evolution. It is likely that many species. questioned how far we have demon-
the origins of each mode lie in one If we return to the larger questions strated the absence or existence of
lineage: Mode 1, an australopithecine?; relating to the evolution of culture culture. On one hand, it may be ar-
Mode 2, H. ergaster; Mode 3, H. from the common ancestor with gued that as all the hominins make
helmei; Modes 4 and 5, Homo sapiens. chimpanzees to modern humans, we and use stone tools, they are culture-
It is clear, however, that these lineages can consider which among the many bearing; on the other hand, some
all diversified into a number of de- species of hominins can be said to might say that as we have no access to
scendent populations that persisted in have possessed culture or, more accu- symbolic thought or language, there is
making the same stone tools. If these rately, how they compared in their no evidence for culture. In other
are species, then speciation was not cultural capacities with either chim- words, the final interpretation de-
the product of any technologically in- panzees or modern humans. The cul- pends on the definition of culture. The
duced development. Indeed, in terms tural capacities of those hominins problem is how to proceed out of the
of evolutionary process, it seems that making Mode 1 alone (early Homo, definitional problem.
technologies change during the course early H. ergaster, and H. erectus) could One way is to recognize that culture
of a lineage’s existence. be seen as very close to that of chim- is neither an absolute, present-or-ab-
The second point is that if the stron- panzees in terms of their limited con- sent trait nor an indivisible whole. It is
gest evidence for the evolution of en- trol and formalization of functional made up of a series of potentialities,
hanced cultural capacities and their output, although the ability to gener- largely resting in cognition, and de-
underlying cognition comes with free- ate standardized stone tools seems to pending on different mental thoughts.
dom from the constraints of the envi- represent some sort of shift (perhaps We can differentiate, for example, be-
ronment (and, in the case of technol- shared with some australopithecines?). tween imitation and copying as one
ogy, this is presumably raw-material Those making Mode 2 (H. ergaster and element, which forms the basis for so-
constraints), then this occurs in a se- H. heidelbergensis) show in the stan- cial transmission, social learning, and
ries of stages during the development dardization of form and the remark- the maintenance of traditions, and in-
of later Mode 2, more fully in Mode 3, able stability of tradition a consider- novation and elaboration, which
and certainly with the elaboration of able difference from the chimpanzee forms the basis for cultural diversifi-

13. Evolutionary Anthropology 121
TABLE 2. Nature and Implications of Changes in Technological Modes
Change through Associated
Transition Nature of change Nature of output time Cultural inferences hominins
Mode 0 Extension from stone- Relatively few different
Little, although the Hominins probably not robustus?
2 tool use to stone- forms, with little Developed very dissimilar from garhi?
Mode 1 tool modification, formal shape Oldowan can be apes, but control habilis
or extension of Regional variation seen as a move and foresight rudolfensis
nonstone tool probably just raw- toward Mode 2 involved in consistent ergaster
modification to material related and greater fracturing, choosing erectus
stone control, but the raw materials, and antecessor
rate of change deploying tools
(100,000s of years) shows a difference
is very slow from the capabilities
of living apes
Mode 1 Ability to strike off Relatively few forms, Considerable The emphasis in Mode ergaster
2 large flake blanks but these show signs change through 2 technology is on heidelbergensis
Mode 2 from cobbles or to of a preferred shape, time that may be greater planning
use large nodules in and often exhibit related both to and goal-directed
ways that allow symmetry technical behavior associated
invasive retouch on Regional variation is competence and with demand for
both sides probably largely the demand for particular shapes
determined by raw particular Some evidence for
materials (flakes preferred shapes cultural variation on
versus nodules) (symmetry) a large
geographical scale
(cleavers in Africa
and India)
Mode 2 Transformation of the Diverse, predetermined Some directional Clear evidence for helmei
2 planning involved flakes, often very change from cultural variants neanderthalensis
Mode 3 in Mode 2 toward thin, with potential early generalized regionally. sapiens
the preparation of for modifying MSA to later, but Evidence for greater
the core to allow extensively into most of the planning and
greater control of different tools interassemblage awareness of
flake production (especially points) variation is the indirect outputs
Regional variation may development of
be increasingly local styles (MTS,
associated with Stillbay,
cultural patterns Chatelperronean)
rather than raw Final Mode 3 in
materials Europe undergoes
major change
Mode 3 Continuation of the Blade blanks for use as Major change Evidence for ethnic sapiens
2 strategy of composite tools and through time, marking by
Mode 4/5 emphasis on flake for secondary although not in technology and
rather than core shaping any unidirectional other elements of
production, and Major regional way toward material culture
predetermination variation goes greater technical Localized cultural
of shape. But with beyond raw material competence or traditions and
emphasis on constraints, and refinement variants are
narrow flakes probably reflects endemic
(blades), very thin active strategies of
flakes, and use and cultural
miniaturization preference
(Mode 5, microliths)
cation. Each of these, and many oth- is that these contrasts are extremes on can track a series of different trajecto-
ers, can be considered as a more finely a continuous scale and, furthermore, ries, each of which contributes to the
graded scale. Whiten, in this volume, that they can vary independently. In final outcome. Thus, although we may
has proposed that culture be tackled this sense, we return to the position never be able to speak in absolute
through the search for contrasting that while culture as the end product terms about the cultural status of ex-
features, which would include such of evolution may be a qualitatively dif- tinct hominins, we may be able to
things as the existence or absence of ferent “whole,” its evolution is best scale them relative to humans and
traditions. Perhaps what the paleoan- treated in a more reductionist and chimpanzees, and also gain insights
thropological perspective adds to this piecemeal manner.11,12 In this way we into the process. For this to occur,