Evolutionary Anthropology 25:6–19 (2016)

ARTICLE

The Acheulean Handaxe: More Like a Bird’s Song Than a Beatles’ Tune?

RAYMOND CORBEY, ADAM JAGICH, KRIST VAESEN, AND MARK COLLARD

The goal of this paper is to provoke debate about the nature of an iconic arti- lion square kilometers, multiple eco- fact—the Acheulean handaxe. Specifically, we want to initiate a conversation logical regions, and roughly a about whether or not they are cultural objects. The vast majority of archeolo- hundred thousand generations. gists assume that the behaviors involved in the production of handaxes were Acheulean handaxes are the defin- acquired by social learning and that handaxes are therefore cultural. We will ing artifact of the Acheulean indus- argue that this assumption is not warranted on the basis of the available evi- try, which also includes flakes, flake dence and that an alternative hypothesis should be given serious consideration. tools, and cores, as well as other This alternative hypothesis is that the form of Acheulean handaxes was at least large cutting tools such as cleavers, partly under genetic control. picks, trihedrals, and unifaces. The Acheulean industry was preceded by the Oldowan, which is found in Named after the site of Saint-Acheul Acheulean handaxes are one of the Africa and parts of Eurasia; it was in France, where they were first identi- commonest, most widely distributed, succeeded by the Middle Palaeolithic fied, Acheulean handaxes are distinc- and longest-lasting archeological arti- in western Eurasia and the Middle tive (Fig. 1). Indeed, they are so facts. Several hundred thousand Stone Age in Africa. distinctive that they are probably the Acheulean handaxes have been recov- Acheulean handaxes are thought to one artifact that all archeologists, ered from sites in many regions of the have been produced by two extinct whatever their period of interest, are Old World, including North, South, hominin species, Homo erectus and capable of identifying. Acheulean han- and East Africa; Europe; and West- Homo heidelbergensis.Fossilsassigned daxes were produced by the bifacial ern, South, and East Asia (Fig. 2). The to H. erectus have been recovered from reduction of a block or large flake oldest Acheulean handaxes date to sites in East Africa, South Africa, blank around a single, long axis. They approximately 1.76 million years ago North Africa, the Caucasus, Southeast have a cutting edge in the secant (Ma)1 and the youngest to between Asia, and East Asia.5 H. erectus is plane, and range in shape from lan- 300 and 200 thousand years ago sometimes subdivided into Homo ceolate through ovate to orbiculate. (Ka).2,3 Thus, they span several mil- ergaster and H. erectus sensu stricto.5

Raymond Corbey is Professor of Epistemology and Anthropology at the Faculty of Archaeology of Leiden University and Associate Professor of Philosophy at Tilburg University, both in the Netherlands. Much of his research deals with the history and philosophy of anthropology and archeology. Email: [email protected].

Adam Jagich is a graduate of Stony Brook University and Leiden University. Currently he is the scientific director of CommonSites, a web- based platform for the heritage sector, and is based in Lagos, Nigeria. His main research interest is Neanderthal biogeography. Email: [email protected].

Krist Vaesen is an assistant professor of philosophy at Eindhoven University of Technology and a research fellow at the Faculty of Archaeol- ogy, University of Leiden, both in the Netherlands. He leads a research program on the Darwinization of the cultural sciences. Email: k.vae- [email protected].

Mark Collard is the Canada Research Chair in Human Evolutionary Studies and Professor of Archaeology at Simon Fraser University in Can- ada. He also holds a personal chair in archeology at the University of Aberdeen in the UK. Collard is an evolutionary anthropologist with a wide range of interests, including the evolution of technology. He is the corresponding author for this paper. Email: [email protected].

Key words: Acheulean handaxe; cultural transmission; social learning; genetic transmission

VC 2016 Wiley Periodicals, Inc. DOI: 10.1002/evan.21467 Published online in Wiley Online Library (wileyonlinelibrary.com). ARTICLE The Acheulean Handaxe: More Like a Bird’s Song Than a Beatles’ Tune? 7

ual signaling.10,11 There is also debate about the extent to which the form (that is, the shape and size) of Acheu- lean handaxes is deliberate. Most researchers assume that it is, but others assert that resharpening greatly affects handaxe form.12 According to this argument, resharpening generates similar forms in assemblages that are geographically and temporally sepa- rated. Still other researchers contend that raw material quality affected han- daxe form,13 with, for example, large, flat chunks of flint from chalk cliffs yielding handaxes of a different form than small, elongated flint pebbles obtained from river beds. There is even debate about the validity of the Acheulean handaxe as a type. The majority of researchers 14 Figure 1. Acheulean handaxes from the site of Boxgrove, England, which dates to about agree with McNabb that while han- 500 Ka. The handaxes are made of flint and are between 12 and 14.5 cm in length. Pho- daxes may be large or small, more or tograph by W. Roebroeks; used with permission. [Color figure can be viewed in the less refined, pointed or ovate, sym- online issue, which is available at wileyonlinelibrary.com.] metrical or off-set, beneath this vari- ability there are general “themes” that are present in all handaxes. However, The former is represented by several mens to H. heidelbergensis and African there are some dissenting opinions. early African specimens and the latter specimens to Homo rhodesiensis. Nicoud,15 for instance, argues that by Eurasian and later African speci- Several issues regarding Acheulean the Acheulean handaxe is not a uni- mens. The hypodigm of H. heidelber- handaxes are contentious. Most con- tary phenomenon but rather an artifi- gensis includes specimens from East spicuously, there is disagreement cial grouping together of artifacts on Africa, South Africa, Europe, South about their function. Most researchers the basis of superficial morphological Asia, and East Asia.6 H. heidelbergensis consider handaxes to be cutting tools, and technological similarities. is sometimes argued to be a but it has also been suggested that Thus, disagreement about Acheu- 7–9 “wastebasket” taxon. Those who they were throwing weapons. In lean handaxes abounds. However, as choose to split this taxon into two spe- addition, it has been proposed that Richerson and Boyd16 observe, there cies usually assign Eurasian speci- they played a role in social and/or sex- is one thing that more or less all researchers working on handaxes agree on, which is that the behaviors necessary to manufacture them were copied from other individuals and, therefore, that handaxes are cultural objects. Richerson and Boyd16 offer an argument against this idea. They point out that both models and ethno- graphic data suggest that cultural learning in the small, relatively iso- lated groups that H. erectus and H. heidelbergensis are thought to have lived in should have resulted in rap- idly diverging traditions rather than the “bewildering”17:648 geographic and temporal stability exhibited by the Acheulean handaxe. Based on this, Richerson and Boyd16 suggest that the conservatism of Acheulean handaxes may be evidence, not of cul- Figure 2. Acheulean handaxes from various regions (to scale; biface 7 is 22.5 cm tall). Sites: 1) Boxgrove, England; 2) North of Bridge Acheulean, near Gesher Benot Ya’aqov, tural transmission, but of genetic 18 Israel; 3) Erg Tihoda€ıne, Algeria; 4) Sterkfontein, South Africa; 5) Olduvai Gorge, Tanzania; transmission. Foley makes a similar 6) Bose, China, 7) Isampur, India.89,124–129 Figure by Shumon Hussain. point, arguing that since the pattern 8 Corbey et al. ARTICLE of handaxe variation does not match of the behaviors required to produce production of handaxes and even sug- what we expect to see if the behaviors handaxes. gest that it was “most likely aided by involved in their manufacture were Cultural transmission has been active teaching.” Bar-Yosef,22 in his socially learned, a role for genetic invoked numerous times in the post- concluding remarks to a widely cited transmission should be considered. 1990 literature to explain the stability edited volume on the Acheulean, sup- To date, the possibility that the pro- of the key features of handaxes over plies another example of the use of duction of Acheulean handaxes was roughly 1.5 million years. Mithen19 cultural transmission to explain the under at least partial genetic control provides a clear example in his 1999 stability of the key features of han- has not been given serious considera- paper, “Imitation and cultural daxes. In a section entitled “The tion by archeologists. In our view, change: a view from the Stone Age, conservatism of the Acheulean this is unfortunate. Here we attempt with specific reference to the manu- knappers,” he argues that the reten- to explain why we think the genetic- facture of handaxes.” Observing the tion of predetermined solutions was transmission hypothesis should be high degree of similarity in form “dictated by cultural concepts.”22:487 treated as a serious contender for among handaxes, Mithen argues that In addition to being invoked to explaining the handaxe phenomenon. imitation must have been a “necessity account for the stability of key fea- We review several lines of evidence in for acquiring the technical skills to tures of the handaxe, cultural trans- 19:393 favor of the hypothesis, as well as evi- manufacture them.” Similarly, mission has been used to explain the dence against the dominant cultural temporal variation in handaxe form. transmission hypothesis. Subse- Over time, handaxes generally quently, we use bird behaviors that became smaller, thinner, and less have a substantial genetic component the possibility that the elongated. They also became more to illustrate how a behavior as com- production of Acheu- standardized and more finely made.23 plex as handaxe production could be Furthermore, some studies have under genetic control. In line with lean handaxes was found that handaxes became more Richerson and Boyd,16 we will use the under at least partial symmetrical as time progressed.23–25 terms “social learning,” “cultural Vaughan4 and Lycett23 used models learning,” and “cultural transmission” genetic control has not from the field of population genetics interchangeably. been given serious con- to account for these changes. Despite sideration by archeolo- borrowing their models from genet- THE DOMINANCE OF THE ics, both authors frame their studies CULTURAL TRANSMISSION gists. In our view, this is in terms of cultural evolution, and therefore implicitly assume that the HYPOTHESIS unfortunate. Here we behaviors involved in the manufac- Until the 1990s, archeologists rarely attempt to explain why ture of handaxes were socially discussed the manner in which the we think the genetic- learned. behaviors involved in the production of Other authors have used cultural Acheulean handaxes were acquired. transmission hypothesis transmission to explain inter- and Handaxes were simply treated as cul- should be treated as a intra-regional differences in handaxes. 26:14 tural objects and the reader was left to serious contender for Mithen, for example, contends infer that this meant the behaviors that “[t]here are clear, well defined involved in their production were explaining the handaxe handaxe types within the Acheulian, learned socially. In the 1990s, this phenomenon. indicating cultural traditions in arte- began to change and, in line with devel- fact form.”1 Later in the same paper opments elsewhere in the discipline of he suggests that intersite differences archeology, researchers working on may relate to “variability in the degree Acheulean handaxes started to pay of experience and opportunities for attention to the transmission of the Shipton, Petraglia, and Paddayya20 social learning by individuals within behaviors involved in their production. state that imitation is crucial for and between hominid groups owing to Significantly, for present purposes, the understanding the conservatism of variation in social interaction and 26:17 27 transmission mechanisms suggested to the Acheulean industry. They contend group size.” Wynn and Tierson, 28 29 be involved in the production of Acheu- that “from the outset Acheulean Roe, and Wenban-Smith make lean handaxes in the vast majority of social interactions were characterized similar arguments. Wynn and Tier- these publications, were cultural. We by a propensity for true imitation and son27 hypothesize that variability have located only a handful of publica- it is this, rather than any genetic pre- among handaxes from Indian, British, tions, among them the aforementioned disposition to produce bifaces, which Israeli, and African sites reflects works by Boyd and Richerson16 and has produced the remarkable homo- regional cultural traditions, while Foley,18 that suggest that something geneity of the Acheulean.”20:229 Roe28 and Wenban-Smith29 both posit other than cultural transmission might Shennan and Steele21:368 agree that cultural transmission of handaxe tra- have been involved in the acquisition imitative learning was crucial in the ditions within the British Palaeolithic. ARTICLE The Acheulean Handaxe: More Like a Bird’s Song Than a Beatles’ Tune? 9

The strength of the belief that a specific function. However, this ongoing debate about the nature of Acheulean handaxes are cultural rebuttal does not take into account the behavioral differences among objects is underscored by a recent copying error, which, because of the wild-living chimpanzee groups. Over paper by Machin.30 She criticizes the limitations of perception, is inevitable the last 40 years, numerous behaviors standard approach to understanding when learning how to make artifacts have been found to vary among chim- handaxe variability, in which a single by copying another individual. panzee groups. Some behaviors occur factor is viewed as being of primary Experiments involving human sub- at some long-term study sites but not importance. Instead, she contends, jects indicate that copies typically at others. For instance, at the site of we need to recognize that multiple deviate from their target by 3%-5%, Bossou, Guinea, individuals often factors would have been involved. and that this error compounds over detach fronds from an oil palm and She goes on to outline no fewer than multiple transmission events.31 Over use them to smash the tree’s crown to 33 48 factors that she believes were the course of almost 1.5 million years, produce a pulp that is edible. So far, likely to have influenced handaxe 3%-5% copying errors would have this behavior has not been docu- 34 variability. Significantly, for present resulted in a substantial amount of mented at any other site. In addi- purposes, at no point does she ques- spatiotemporal variation, regardless tion, the way in which certain tion the assumption that the behav- of how closely linked handaxes were behaviors are performed varies iors required to produce handaxes to a particular function. This is not among sites. Nut cracking illustrates were acquired by cultural transmis- just a theoretical argument. A recent this. At Bossou, only stone hammers study found that size variation in a and anvils are used to perform this sion. The fact that even an author 35 who calls on colleagues to think large sample of Acheulean handaxes task, whereas in the Ta€ı Forest, Cote^ much more broadly about the way in was lower than expected from copy- d’Ivoire, two different kinds of ham- 32 which they approach handaxes does ing error. Given that copying error mer, wood and stone, and two differ- not question the assumption that would inevitably have occurred if ent kinds of anvil, root and stone, are used to crack nuts.36 An intersite handaxes are cultural objects shows individuals had acquired the behav- comparison carried out a few years just how deeply embedded the iors required to produce handaxes via ago suggested that at least 65 behav- assumption is in archeology. social learning, this supports the idea that handaxes were not fully cultural iors vary among chimpanzee groups.37,38 REASONS TO DOUBT THE artifacts and may have been under at least partial genetic control. The nature of the among-group CULTURAL TRANSMISSION As strong as we think Richerson behavioral differences is contested. HYPOTHESIS and Boyd’s16 and Foley’s18 argument According to some researchers, many of the behaviors are likely to be socially Earlier we explained why Richerson is, we are of the opinion that it is learned and therefore meet the main and Boyd16 and Foley18 suggest that a just one of the problems facing the condition that is necessary for being role for genetic transmission in the cultural transmission hypothesis. We recognized as cultural behaviors.34–41 manufacture of handaxes should be will outline several other reasons to question the assumption that the Other researchers question whether the considered. To reiterate, they point behaviors are socially learned42–46 and out that the long-term conservatism of behaviors involved in the manufac- ture of handaxes were acquired highlight the fact that many of the puta- the Acheulean handaxe is inconsistent tive cultural behaviors occur in a single purely by cultural learning and to with models of cultural evolution and subspecies. They point out that genetic take seriously the possibility that ethnographic data, which suggest that studies suggest that some chimpanzee they were acquired at least in part cultural transmission in groups of the subspecies have been isolated for hun- by genetic transmission. We believe type H. erectus and H. heidelbergensis dreds of thousands of years and con- that, taken together, Richerson and are thought to have lived in should tend that in these circumstances it is Boyd’s16 and Foley’s18 argument and give rise to considerable spatiotempo- not possible to discount a genetic origin the evidence we discuss in this sec- ral variability. That handaxes do not for the behavioral differences. tion represent a powerful case exhibit the kind of signal predicted by Currently, it is unclear which of against continuation of the current, cultural evolutionary models and eth- these hypotheses is correct. Both have 16 uncritical approach to the nature of nographic data, Richerson and Boyd been supported by recent work. Lycett, 18 the Acheulean handaxe. and Foley aver, implies that han- Collard, and McGrew47–49 have daxes are not fully cultural objects, reported several studies in which they which in turn suggests that the behav- The Chimpanzee Culture compared phylogenetic trees derived iors involved in their manufacture Debate Demonstrates That from chimpanzee behavioral datasets may have been acquired at least in Mechanisms of Transmission to published genetic data pertaining to part by genetic transmission. Need to be Investigated and the relationships among chimpanzee The most obvious objection to subspecies and found that the behav- 16 18 Not Simply Assumed Richerson and Boyd’s and Foley’s ioral trees did not match the genetic argument is that it ignores the possi- Another reason to question the data. Based on this, these authors47–49 bility that the form of handaxes was assumption that Acheulean handaxes conclude that the culture hypothesis is stable because it was strongly tied to are cultural objects is provided by the more likely to be correct than is the 10 Corbey et al. ARTICLE genetic one. Langergraber and col- without any attempt to consider learning, but there was some effect leagues46 have reached a different con- alternative mechanisms of transmis- of social learning. Variation in pref- clusion. Using the Mantel test, they sion. In 2004, McElreath52 reported erence for friends or kin was not found a significant correlation between a ground-breaking study in which he influenced at all by social learning, genetic and behavioral data for several sought to identify the processes but only by individual learning. The chimpanzee groups. Based on this underlying differences in the beliefs main significance of these results for they argue that it is not possible to rule and behavior of farmers and pastor- the present argument is that they out a major role for genes in generating alists living in the Usangu Plains of demonstrate that even the beliefs the differences among the groups’ Tanzania. He focused on individuals and behaviors of modern humans behavioral repertoires. There is dis- who had changed household econo- cannot be simply assumed to be cul- agreement about why the results of tural phenomena. Many of them these studies do not converge,50,51 probably are, but some of them which means that, at the moment, it is likely are not. As such, analysis is not possible to state with confidence Most researchers have required before a given belief or that the behavioral differences among simply assumed that the behavior is deemed cultural. chimpanzee groups are cultural or To put it another way, McEl- genetic. behaviors involved in reath’s52 findings call into question Thus, the nature of the behavioral the manufacture of han- the anthropological axiom that all differences among chimpanzee groups the interesting beliefs and behaviors remains unclear even after nearly 40 daxes were acquired by of modern humans are cultural. By years of discussion and completion of cultural transmission. As extension, they also challenge the several research projects specifically default assumption of paleoanthro- designed to resolve the issue. This far as we are aware, no pologists that the behaviors of clearly indicates that it is not straight- attempt has ever been extinct hominins that are reflected in forward to infer mechanisms of trans- made to determine the archeological record, including mission from behavioral patterns. An the behaviors that resulted in Acheu- obvious corollary of this is that we whether the available lean handaxes, are cultural. Many of should be skeptical when a mecha- data support the cul- the behaviors of modern humans nism of transmission has been con- likely are cultural, and many of the cluded to be responsible for a tural- transmission behaviors of extinct hominins prob- behavioral pattern without any hypothesis better than ably were cultural. But it is not attempt to consider alternative mecha- defensible to simply assume that this nisms of transmission. This is the case they support potentially is the case. Rather, it needs to be with the Acheulean handaxe. Most competing hypotheses demonstrated analytically. We need researchers have simply assumed that to treat the processes underlying the the behaviors involved in the manufac- such as individual behaviors that anthropologists and ture of handaxes were acquired by cul- learning and genetic archeologists study as unknown and tural transmission. As far as we are transmission. design research projects to test aware, no attempt has ever been made between the competing hypotheses, to determine whether the available in the same way that those working data support the cultural transmission on the processes underlying behav- hypothesis better than they support ioral variation among chimpanzee potentially competing hypotheses groups have tested between the cul- such as individual learning and ture hypothesis and the genetic mies, and investigated whether their genetic transmission. Accordingly, the hypothesis. Given the difficulty of dispositions with respect to three basis for believing that handaxes are associating artifacts with hominin domains — preference for friends or cultural objects is shaky. species and the fragility of DNA, kin, respect for elders, and belief in doing so will be difficult. However, witchcraft — were a consequence of this is not a good reason to ignore Work on Modern Humans Also individual or social learning. McEl- the problem. Indicates That Mechanisms of reath’s52 results were mixed. He Transmission Need To Be found that variation in the preva- The Failure of Acheulean Investigated and Not Simply lence of belief in witchcraft was best explained by social learning. How- Handaxes To Track Assumed ever, variation in the other two Environmental Variation Is Recent research on living humans domains was better explained by Inconsistent with Models of also suggests that we should be skep- individual learning. Variation in Cultural Evolution tical when a mechanism of transmis- whether or not individuals believed sion has been assumed to be that respect for elders is important As we have seen, the long-term responsible for a behavioral pattern was strongly influenced by individual conservatism of the Acheulean ARTICLE The Acheulean Handaxe: More Like a Bird’s Song Than a Beatles’ Tune? 11 handaxe is inconsistent with a key track environmental variability presents cognitive ability would be denied to prediction of recent models of cul- another substantial challenge to the cul- Neanderthals, despite compelling tural evolution, namely that cultural tural transmission hypothesis. arguments to the contrary.64–66 entities should exhibit considerable The second potential explanation spatiotemporal variability. Handaxes Explaining the Increased Speed for the increase in speed of artifact are inconsistent with models of cul- of Cultural Change in The Late change at 300–200 ka concentrates tural evolution in other ways. on demography. Hominin population Modeling work carried out over Pleistocene Is Difficult If size is thought to have increased in the last 30 years or so sets fairly Acheulean Handaxes Are the Late Pleistocene and it has been stringent conditions on the emer- Assumed To Be Fully Cultural argued that this would have given gence of cultural transmission.53,54 rise to an increase in cultural com- A further problem facing the cul- The goal of this work was to deter- plexity. The main intuition behind tural transmission hypothesis is the mine when cultural learning is the dependence of cultural change much more rapid pace of change favored over individual learning and on demography is that larger popula- after Acheulean handaxes disap- genetic transmission, given that cul- tions can sustain a more complex peared at 300-200 Ka. If the behav- tural learning is cognitively costly. culture because they are more likely iors involved in the production of For present purposes, the key finding to contain individuals whose skill both handaxes and post-Acheulean is that cultural learning evolves only level is at least as good as that of the artifacts were culturally learned, how in moderately variable environments. individuals in the previous genera- do we explain this very marked In slowly changing environments, tion. However, the demographic increase in the speed of change? genetic transmission outperforms explanation of the transition toward Currently, two hypotheses prevail. cultural transmission, while in highly fully modern human behavior is variable environments individual Unfortunately, neither is particularly problematic in at least two respects. learning beats both genetic transmis- convincing. First, it has been shown that the for- sion and cultural transmission. The first hypothesis focuses on cog- mal model that is used most fre- Equally importantly, in order for cul- nition. According to this hypothesis, quently to support the demography tural transmission to be favored over the shift in speed occurred because hypothesis67,68 only predicts a rela- the alternatives it must result in the hominin brain was “upgraded”; tionship between population size behaviors that track environmental that is, one or more cognitive capaci- and cultural complexity in implausi- conditions. If it does not give rise to ties, such as enhanced working mem- ble conditions.69,70 Second, a key ory, causal reasoning, or executive such behaviors, genetic transmission 59–63 prediction of the hypothesis — that or individual learning will be pre- control, were added. These herit- there should be a significant, positive ferred by selection, depending on the able biological changes introduced relationship between population speed of environmental change. the necessary cultural variation, so size and cultural complexity in For the behavioral routines involved that the cultural evolutionary process hunter-gatherers — is not supported in the production of handaxes to have could yield more complex culture. empirically. Several studies have been culturally transmitted, therefore, This scenario faces several problems, found no evidence of an impact of the environments in which handaxes depending on when one sets the date population size on technological were produced must have been variable of the event. If the event is set around richness and complexity in and handaxes must have tracked that 100 Ka, one can account for the spo- ethnographically documented variability. The first condition is met. It radic appearance of many markers of hunter-gatherers.71–74 A similar find- is clear that Acheulean toolmakers had modern behavior at multiple African ing was obtained in a recent attempt to cope with fluctuating climates.55–58 sites as early as 70 to 90 ka. However, to use North American archeological They also expanded into various new one cannot explain the disappearance data to evaluate the impact of popu- habitats.55,56 Thus, there is reason to of these markers between 75-60 Ka lation size on hunter-gatherer think that the environments in which and their later reemergence at differ- technology.75 Moreover, results of handaxes were produced were variable. ent times in various parts of the attempts to estimate the population In contrast, the second condition is not world. If the event is set later, say sizes of Late Pleistocene human pop- met. As we explained earlier, the key right before the transition in Europe ulations conflict with the timing of features of the handaxe do not vary (45 Ka), one needs to assume that the the transition toward fully modern much across space or through time. arrival of fully modern behavior in human behavior.76 The implication of this is that handaxes southeast Asia and Australia resulted Given that neither of the cultural didnotrespondtotheenvironmental from either convergent evolution or hypotheses withstands scrutiny, how variability that Acheulean toolmakers dispersal out of Europe. The former do we explain the shift in speed of experienced. Given that cultural evolu- seems unlikely, given the drastic artifact change 300-200 ka? If Acheu- tionary models suggest the ability to nature of the change, while the latter lean handaxe production behavior track environmental variability is the is hard to reconcile with the available was under genetic control, the post- reason why cultural transmission is evidence on Late Pleistocene migra- Acheulean speeding up and diversifi- sometimes favored over genetic trans- tions. Furthermore, unless conver- cation of culture perhaps marked mission, the fact that handaxes do not gent evolution is invoked, increased another event, namely greater 12 Corbey et al. ARTICLE reliance on cultural inheritance. cooperation.79–82 may be another fac- river basins. Whenever this hap- This, in turn, may imply either the tor facilitating the ontogenetic devel- pened, so the hypothesis goes, the evolution of special-purpose mecha- opment of cultural learning. In hominins occupying the area pro- nisms for cultural learning or collective rituals, groups express and duced bifaces. On the face of it, this domain-general mechanisms for reaffirm shared beliefs, norms, and seems like a reasonable suggestion. learning being “tuned,” ontogeneti- values, and this enhances communal However, it actually ignores a crit- cally, to inputs from other agents.77 stability and group harmony.83,84 In ically important question: How were The first option seems to suffer addition, belief in supernatural enti- the behaviors involved in the produc- from the same problems as those ties that monitor their devotees and tion of handaxes acquired by homi- noted in relation to the cognitive punish them in case of transgressions nins located 1,500 km from the upgrade hypothesis. The second has been proposed to encourage pro- Movius Line? option is more promising because it sociality towards strangers.85,86 Reli- There are two potential answers to leaves room for cultural differences gion would thus seem to provide this question that involve cultural within Homo sapiens (geographically suitable conditions for cultural learn- transmission, neither of which is and diachronically) and between H. ing and therefore allow the formation particularly compelling. One possi- sapiens and Neanderthals, as a result of cumulative culture. bility is that the Chinese hominins of differences in ontogeny. According To summarize, the increase in the maintained contact with groups that to this hypothesis, the tempo of cul- pace of artifact change after the regularly produced handaxes far to tural evolution depends on the extent Acheulean may not be the conse- the west. The other is that Chinese to which attentional, perceptual, quence of a cognitive upgrade or hominins transmitted the knowledge and/or motivational mechanisms are increased population size, as archeolo- of handaxe manufacture across ontogenetically biased toward others. gists often suppose. Instead, it may many generations without ever pro- The Late Pleistocene transition may indicate a shift from genetic transmis- ducing them. Given what little we be a consequence of hominins start- sion to increasing reliance on ontoge- know about population dynamics, ing to rely on cultural learning for a netic cultural learning. This proposal life history traits, and patterns of dis- wide range of behaviors, due, for offers a plausible mechanism by tribution so deep in the past both instance, to local conditions favoring which artifact change could acceler- scenarios seem unlikely, especially in increased social tolerance and/or the ate. It also is easier to reconcile with view of the vulnerability of cultural emergence of new types of social geographic and temporal variation transmission to information loss.90 organization in which contact than is the cultural hypothesis. In contrast, the genetic transmis- between putative pupils and models sion hypothesis straightforwardly is more likely. explains the occurrence of handaxes With regard to the latter possibility, Some Geographic Facts About 1,500 km from the Movius line: The a recent study by Hill and Handaxes Are More Consistent genes responsible for the mainte- coworkers78 indicates that it is com- With Genetic Transmission Than nance of such behaviors remained mon for contemporary hunter- With Cultural Transmission unexpressed in the population until gatherers to live in residential units 87,88 with a high percentage of unrelated In the late 1940s, Movius drew proper raw material became avail- individuals. Such organization attention to the geographic distribu- able, at which time the relevant implies friendly visits between units tion of Acheulean handaxes. He genes were expressed. The feasibility and many more opportunities for pointed out that handaxes had been of this scenario is illustrated by stud- observing novel cultural traits than is found in Africa, Europe, and India, ies showing that organisms are able the case when residential units are but not in East and Southeast Asia. to respond to changing environmen- primarily kin-based. The emergence This “Movius Line” persisted for tal conditions by alterations to gene 91 of these sorts of networks may thus many years, but recently handaxes expression. be one of the ways in which human have been found at various sites in A fourth potential explanation for learning became more attuned to the Bose Basin and other river the Chinese handaxes is worth noting. social cues. Also, if group members basins in China, some 1,500 km It is possible that the handaxes are 89 are not closely related, one should northeast of the Movius Line. This the result of convergent evolution; expect a variety of mechanisms that unexpected development presents a that is, hominins in China developed actively promote cooperation (indi- third major problem for the cultural handaxes independently from homi- rect reciprocity, sanctioning of defec- transmission hypothesis. nins on the west side of the Movius tors). Under these conditions, it The unexpected occurrence of han- Line, so that the Chinese handaxes seems likely that social stimuli pro- daxes in China has been explained in are not, in fact, Acheulean handaxes. vide reliable information, which, terms of raw material distribution. It This is possible; over the last few dec- according to Heyes,77 in turn favors has been suggested that high-quality ades, it has become clear that conver- higher attentiveness to such stimuli. lithic raw materials are rare in the gent evolution is commonplace.92–94 Religious beliefs and practices, eastern part of the range of H. erec- However, the notion that the Chinese which have been suggested to pro- tus, but forest fires periodically handaxes and Acheulean handaxes mote intragroup cohesiveness and uncovered suitable raw material in were invented independently is a less ARTICLE The Acheulean Handaxe: More Like a Bird’s Song Than a Beatles’ Tune? 13 parsimonious explanation than the through observation would become behaviors. We begin by discussing gene expression-based explanation. pre-programmed (‘innate’) to some bird song and bird tool-use as “soft” There is another geographic fact degree during phylogenetic histo- options that involve a combination about handaxes that is illuminating ry.”97:351 They go on to review empir- of genetic influence and social learn- to consider from the perspective of ical evidence suggesting that ing. We then discuss a “hard” option the genetic transmission hypothesis, Hebbian learning, in which simulta- — structure building by birds, in and that is the way in which variabil- neous activation of neurons leads to which strong genetic determination ity in symmetry changes with dis- increased interconnection between interacts with the availability of raw tance from Africa. Using a model those neurons, and genetic preprog- material and local environmental derived from population genetics, ramming occur specifically in con- conditions but not with much, if Lycett23 tested aspects of handaxe nection with manual action. any, cultural learning. symmetry in relation to predictions Another relevant insight from neu- derived from the neutral theory of roscience is that the neuronal cir- Bird Song and Tool-Use evolution. His analyses indicated cuitry in the human brain Bird song, a well-studied behavior, that the intra-assemblage variability responsible for language production is strongly influenced by genetic pre- of handaxe symmetry decreases with not only overlaps considerably with dispositions and implemented by dis- distance from Africa and does so in areas controlling manipulation but crete, well-defined neural circuits.101 a way that is inconsistent with itera- also is under strong genetic control. 98,99 Songbirds, hummingbirds, and par- tive founder effect. On the basis of Stout and colleagues argue that rots have fixed, species-wide basic this, Lycett23 argues that aspects of both tool-making and speech are songs from which individuals and handaxe symmetry were subject to multilevel, hierarchically nested, regional subpopulations develop selective forces. Not surprisingly, sequential, and goal-directed motor their own varieties. Auditory-guided given the strength of belief in the sequences. They summarize their vocal motor learning is important, field that handaxes are cultural argument as follows: “The observed for when members of these taxa are objects, Lycett,23 throughout his patterns of activation and of overlap raised in isolation their songs remain paper, assumes that the behaviors with language circuits suggest that simple. The genetically controlled involved in handaxe manufacture tool-making and language share a early song of zebra finches (Taenio- were culturally learned. Lycett men- basis in more general human capaci- pygia guttata), for instance, is plastic. tions the possibility that genetics ties for complex, goal-directed action. The young need several months of influenced the design of handaxes, The results are consistent with co- evolutionary hypotheses linking the social learning to perfect their indi- but does not take it seriously. We 102 vidual, adult song. respectfully suggest that this is a emergence of language, tool-making, population-level functional lateraliza- It appears that in some bird spe- mistake. If the variation in handaxe cies, tool use also combines fixed symmetry fits population genetics tion and association cortex expansion 99 species-wide basic motor behaviors models, it stands to reason that the in human evolution.” We think this line of enquiry is promising with with cultural learning. It has been behaviors involved in the production respect to a possible genetic compo- found, for example, that hand-raised of handaxes may have been acquired nent in handaxe design, but suspect Caledonian crows (Corvus monedu- by genetic transmission. that the capacities in question may loides) reliably develop leaf-tool well turn out to be much more spe- manufacture without ever having Recent Evidence from cific than Stout et al. suggest. Rele- observed it in others, but never Neuroscience Is Consistent with vant points of comparison between develop the sophisticated behaviors the Genetic Transmission tool-making and speech are manual found in wild populations where Hypothesis control and articulatory control, individuals are exposed to models action sequences and syntax, as well and competitors. A study by Ken- 103 Neuroscience has shown that as mental templates and semantics, ward and coworkers illustrates present-day humans can gauge with bird song as a promising third this. By hand-rearing several crows intentions and feelings and can copy relevant field.100 in isolation, they showed that the and predict the actions of others via tool behavior of isolated individuals mirror neurons.95,96 At first sight, BIRD MODELS FOR HANDAXE was so similar to that in wild ones this seems to support the current that it was difficult to avoid the con- approach to Acheulean handaxes in MANUFACTURE clusion that the behavior is innate. terms of domain-general social learn- An obvious concern about the They argue against “the extreme pos- ing. However, studies have begun to genetic transmission hypothesis is sibilities that tool-use depends link mirror neurons with domain- whether the production of an object entirely on social inputs (i.e., is sus- specific genetic preprogramming. as complex as an Acheulean handaxe tained exclusively by cultural trans- Del Giudice, Manera, and Keysers97 could really be under genetic control. mission and thus does not reflect a argue that it is “reasonable to expect We contend that the idea is not dedicated evolved adaptation), and that an ability as crucial for survival implausible if handaxe production is that it has a purely individual, as action recognition and learning compared to some complex bird insight-based origin.”103:1340 If that 14 Corbey et al. ARTICLE

insight, but not necessarily complex, overall planning, conscious decision- making, or a mental image of the ulti- mate goal. The nest is an emergent property of the stimulus-response chain. The bower-birds (Ptilonorhynchi- dae) of New Guinea provide another excellent set of examples. These birds create hut-like shelters and decorate their interiors and the area immedi- ately in front of them with brightly colored objects (Fig. 4). These “bowers” are individually and ecologi- cally variable as a result of individual learning and, possibly, some cultural Figure 3. Nests created by long-tailed tits (Aegithalos caudatus), Great Britain. Long- learning. It is thought, however, that tailed tits build their nests from thousands of pieces of lichen, moss, and spiderweb, as the stability of their basic form indi- well as numerous feathers. The nests are thought to be a result of innate instructions and cates that genes underlie their design individual learning. Photograph A shows a nest under construction, while photograph B 109 shows a completed nest. (Photograph A: Alan Shearman/Wikimedia Commons; Photo- and construction. graph B: nottsexminer/Wikimedia Commons). [Color figure can be viewed in the online As with handaxes, the bowers issue, which is available at wileyonlinelibrary.com.] show a mix of variation and uni- formity. Individual bowers display variation, but share a general form within a species. The same can be were the case, Kenward and col- Structure Building by Birds said for handaxes: Beneath their var- leagues103:1340 contend, we would iability, general themes are present It appears that one important reason not see “inherited action patterns in all of them. why archeologists tend to think that that must have evolved through Structure building by birds is also the behaviors involved in the produc- selection and that are crucial in sus- heuristically relevant when trying to tion of handaxes were socially learned taining tool-oriented behavior in clarify exactly what was under genetic is that the handaxe chaine operatoire adult crows.” The same point has control in handaxe manufacture. It required multiple decisions at multiple been made with respect to the tool- probably was not just a simple target use behavior of hyacinth macaws, stages, from the selection of the size, form, but rather a predisposition shape, and quality of raw material to Egyptian vultures, and woodpecker 15 toward the basic behavioral routines finches, the only other bird species the final retouch. Again, this argu- involved, such as invasive bifacial that is known to habitually use stick ment seems reasonable on its face. reduction while realizing cutting tools in the wild.104–106 However, it is not as secure as it first edges in the secant plane, working Thus, if modeled on bird song and appears. The reason for this is that the from the tip down, and keeping sym- bird tool-use, the production of construction of nests and other struc- metry. These routines would have Acheulean handaxes would have tures by birds often involves long hier- operated in combination with causal involved both genetic transmission archical sequences even though, as is understanding, manipulative skill, and social learning. Raw material generally agreed, the required behav- and intuitive (“folk”) physics.110 selection, the manufacturing process, iors do not involve much, if any, cul- Although Gowlett111:217 thinks of han- and basic design principles would tural learning. daxes as cultural objects, our pro- have been under genetic control, but Eurasian long-tailed tits (Aegithalos posal is consistent with his approach fine-tuned through social learning. caudatus) exemplify the potential when he argues against the notion of The latter would have been depend- complexity of nest building behavior. a template as being “hard and fast” ent on the presence of role models These birds create intricate spherical and instead talks of “instruction sets during sensitive periods, but would nests from thousands of pieces of [from which] a computer could gener- not necessarily have involved cul- lichen, moss, and spiderweb, as well ate the form of a biface mindlessly.” tural group-specific templates or as numerous feathers (Fig. 3). The Unfortunately, genetic evolution usu- explicit instruction. The combination final product is accomplished through ally happens too slowly for researchers of genetic transmission and social a combination of innate instructions to adequately study changes in geneti- learning is predicted to produce uni- and individual learning.107,108 The for- cally controlled behaviors such as nest formity of overall design (due to the mer specifies materials and methods building by birds. The time needed to fixed component) and slight local and a limited repertoire of repetitive, observe changes in evolutionary trends variance (due to the process of stereotyped actions. There is a chain extends far beyond the current length socialization), which is the pattern of stimuli and responses that presup- of study of such phenomena. Direct observed with Acheulean handaxes. poses construction rules and local comparisons between the handaxe and ARTICLE The Acheulean Handaxe: More Like a Bird’s Song Than a Beatles’ Tune? 15

to be a particularly good candidate for another large cutting tool that was under at least partial genetic control. Cleavers are bifacially reduced pieces with a straight cut- ting edge perpendicular to the long axis of the piece. They were mostly made on flake blanks struck from bifacially prepared, usually more or less tortoise-shaped cores. The clea- ver has been argued to be a subcate- gory of the Acheulean handaxe. If this is the case, then parsimony sug- gests they could have been under partial genetic control as well. We suspect the situation is differ- ent with respect to the handaxe-like tools in the industries that came after the Acheulean. As we explained earlier, in Europe the Acheulean was replaced by the industries of the Middle Paleolithic between 300 and 200 Ka. Several European Late Mid- dle Paleolithic industries contain bifacial tools, including the French Mousterian of Acheulean Tradition and the Middle European Keilmes- sergruppen, also called the Mico- quian.112,113 The tools in question resemble Acheulean handaxes, but we think it is unlikely that they were under genetic control. There are two reasons for this. Figure 4. Bowers of the Vogelkop bowerbird (Amblyornis inornatus), western New Guinea. First, Late Middle Paleolithic bifaces It is thought that the within-species stability of the basic form of these bowers indicates are actually only superficially similar that genes underlie their design and construction. Note the organization of colorful and to Acheulean handaxes. A recent shiny items in the area in front of the bower in photograph A. In photograph B, note the way in which the sticks that form the bower are intricately woven together. (Photograph morphometric comparison of Acheu- A: Tim Laman/naturepl.com; photograph B: Ingo Arndt/naturepl.com). [Color figure can lean and Late Middle Palaeolithic be viewed in the online issue, which is available at wileyonlinelibrary.com.] handaxes from Western Europe done by Iovita and McPherron12:69 demon- strated that Acheulean handaxes show an allometric pattern consist- animal behaviors such as birds’ tool Were these tools also partly geneti- ent with reduction of the tip, while use or structure building are therefore cally determined or were they solely Late Middle Palaeolithic handaxes difficult. Nevertheless, mounting evi- cultural phenomena? “show a pattern which implies a dence points to a central role for genes McElreath’s52 discovery that in his maintenance of shape throughout in maintaining aspects of animal tech- study-groups different mechanisms the reduction continuum.” The nology. We suggest that the same underpinned attitudes toward friends authors conclude from this that the should be considered for Acheulean versus kin, respect for elders, and two tool types are unrelated. In addi- handaxes. belief in witchcraft clearly indicates tion, Late Middle Paleolithic bifaces that there is no reason to expect all seem to have been used in a different THE NATURE OF OTHER Acheulean tools to involve the same way than Acheulean handaxes. The transmission mechanisms. Neverthe- latter were primarily platforms on HANDAXE-LIKE TOOLS less, we are of the opinion that the which various working edges (for It is important, under the genetic applicability of the genetic transmis- example, a point, a notch, or cutting transmission hypothesis, to clarify sion hypothesis should be investi- edge) were made, whereas Middle the nature of the other large cutting gated in connection with the other Paleolithic bifaces usually served as tools in the Acheulean industry and large cutting tools of the Acheulean, complete tools with well integrated the handaxe-like bifaces in the indus- among them cleavers, picks, trihe- working edges that remained inte- tries that come after the Acheulean. drals, and unifaces. Cleavers appear grated during resharpening.12,112,113 16 Corbey et al. ARTICLE

The other reason we think it limited variability in key aspects of handaxes are treated as cultural unlikely that Late Middle Paleolithic form over roughly 1.5 million years objects. bifaces were under genetic control is across two continents, in almost The cultural transmission hypothe- that the industries they were part of every type of ecology occupied by sis, however, faces several problems. were geographically restricted and archaic hominins — is almost unani- It underestimates the difficulty of relatively short lived. Ruebens114 has mously explained by cultural learn- inferring transmission mechanisms recently reported a typo-technological ing, variously combined with such from artifacts. It conflicts with cul- analysis of bifacial tools from late factors as individual learning, func- tural evolutionary models and ethno- Middle Palaeolithic sites in several tional requirements, raw material graphic data on key characteristics European countries, including Brit- constraints, similar ecology, reuse of cultural objects. Moreover, this ain, Belgium, the Netherlands, France and resharpening, and drift. In light hypothesis is difficult to square with and Germany. She found evidence of of this apparent consensus, one the greatly increased rate of techno- the existence of three different would expect that the status of han- logical change observed in post- regional entities, the Mousterian of daxes as cultural objects was settled Acheulean industries. It also fails to Acheulean Tradition, the Keilmesser- account adequately for the handaxes gruppen, and a new entity, which she that have recently been found calls “the Mousterian with Bifacial beyond the Movius Line. We have Tools.” The Mousterian of Acheulean there are two strands of argued that these and other prob- Tradition, located in southwest lems can be better accommodated by Europe, is dominated by handaxes. positive evidence for the what we call the genetic transmis- The Keilmessergruppen is found in genetic hypothesis. First, sion hypothesis, which is that han- northeast Europe and is distin- daxe production was passed between guished by backed and leaf-shaped handaxe variation generations at least in part through bifacial tools. The Mousterian with accords with predictions genetic inheritance. In addition, Bifacial Tools is located between the there are two strands of positive evi- Mousterian of Acheulean Tradition from models of genetic dence for the genetic transmission and the Keilmessergruppen and con- evolution; second, neu- hypothesis. First, handaxe variation tains a wider variety of bifacial tools accords with predictions from mod- than either of them. That these and roscientific findings sug- els of genetic evolution. Second, neu- other Late Middle Paleolithic indus- gest that simple tool use roscientific findings suggest that tries were geographically restricted may be the result of simple tool use may be the result of and short-lived means they match domain-specific genetic what we expect to see if the behaviors domain-specific genetic preprogramming. involved in their manufacture were In this paper, we have adopted preprogramming. 16,53,115,116 socially learned.16,18 The corollary of dual-inheritance theory’s this is that Late Middle Paleolithic distinction between two channels of bifaces are likely purely cultural inheritance, genetic and cultural, objects. even though, in reality, traits may Based on the results of their morpho- very well resist such neat partition- 16 metric analysis, Iovita and McPher- through careful study a long time ing. We recognize that in doing so ron12 argue that Acheulean handaxes ago and is backed by a substantial we have opened ourselves up to the and Middle Paleolithic bifaces reflect body of evidence. This is not the charge of simply replacing one sim- “different and likely unrelated cultural case, however. The assumption that plistic account with another simplis- and human evolutionary contexts.” We cultural transmission must have tic implausible story. Even though suggest a different scenario in which been the mechanism responsible for this may turn out to be correct, we genetic transmission stopped being a the conservation of specific design hope that setting up the issue in factor in the behavioral routines characteristic of the handaxe can be terms of a dichotomy will spark a involved in the production of Middle traced back to the early nineteenth debate on what has been unquestion- Paleolithic bifaces. How and why this century, when they were first discov- ably assumed by the vast majority of happened is an intriguing question but, ered in France. In the absence of archeologists for more than a given the connection between cultural concepts like genes or instinct, han- century. learning and environmental variability daxes were automatically perceived It may be helpful to clarify what discussed earlier, a link with an as cultural products, similar to those the genetic transmission hypothesis increase in environmental variability of that period’s booming industrial does and does not entail. First, it seems likely.55–58 technology. Crucially, the cultural does not imply strong instruction- transmission hypothesis has gone ism, which is the idea that genes untested ever since. Handaxe pro- code for handaxe production in the CONCLUSIONS duction was, and still is, assumed by sense of being uniquely responsible The remarkably conservative char- most archeologists to be under- for specifying that behavior.117 acter of the Acheulean handaxe — pinned by social learning. Therefore, Developmental factors may well have ARTICLE The Acheulean Handaxe: More Like a Bird’s Song Than a Beatles’ Tune? 17 played an essential role in handaxe from being genetically controlled to ioural contexts of Pleistocene Homo sapiens from Middle Awash, Ethiopia. Nature 423:747– manufacture. Acheulean toolmakers, purely cultural over the course of 752. for instance, would have inherited nearly three-and-a-half million years. 4 Vaughan CD. 2001. A million years of style from other individuals niches con- But it is also possible that stone-tool and function: regional and temporal variation taining stimuli related to handaxe technology went from purely cultural in Acheulean handaxes. In: Hurt TD and Rakita GFM, editors. Style and function: conceptual production. Such niches would have to genetically controlled to purely issues in evolutionary archaeology. Westport, provided cues for individual learn- cultural again. Indeed, based on CT: Bergin and Garvey. p 141–163. ing. Their layout would also have McElreath’s52 results, we need to 5 Hublin JJ. 2014. Paleoanthropology: Homo encouraged the task to be structured erectus and the limits of a paleontological spe- allow for an even more complicated cies. Curr Biol 24:R82–R84. in very specific, previously tested scenario in which the three indus- 6 Stringer C. 2012. The status of Homo heidel- 118 ways. tries have different combinations of bergensis (Schoetensack 1908). Evol Anthropol In accordance with much other transmission mechanisms. 21:101–107. work on cultural evolution, we have Finally it is worth noting that 7 O’Brien E. 1981. The projectile capabilities of an Acheulian handaxe from Olorgesailie. Curr “black-boxed” specific mechanisms when nonhuman animals display Anthropol 22:76–9. driving the suggested genetic evolu- complex behavior, the default 8 Calvin WH. 1993. The unitary hypothesis: a tion of handaxes, such as stabilizing assumption is that it is under genetic common neural circuitry for novel manipula- or directional selection, genetic drift, tions, language, plan-ahead and throwing. In: control. For complex behavior in Gibson KR, Ingold T, editors. Tools, language genetic assimilation, and Baldwinian humans and other hominins, how- and cognition in human evolution. Cambridge: 119 evolution. However, like Sterelny ever, the default position is to invoke Cambridge University Press. p 230–50. 120 and others, we see the Baldwin culture and not genes. This double 9 Samson DR. 2006. Stones of contention: the Acheulean handaxe lethal projectile contro- effect as a plausible candidate. standard betrays a form of anthropo- versy. Lithic Technol 31:127–135. Learned responses to environmental centrism. We contend that the 10 Gamble C. 1997. Handaxes and Palaeolithic stimuli can be replaced by genetically genetic transmission hypothesis individuals. In: Ashton N, Healey F, Pettitt P, controlled behaviors that occur with editors. Stone age archaeology. Oxford: Oxbow deserves serious consideration for Books. p 105–109. little or no learning as a result of 123 this reason as well. 11 Kohn M, Mithen S. 1999. Handaxes: prod- selection pressure on the technical ucts of sexual selection? Antiquity 73:518–526. abilities of the species. Accumulation ACKNOWLEDGMENTS 12 Iovita R, McPherron SP. 2011. The handaxe of technology-related genes “steering” reloaded: a morphometric reassessment of We thank Rob Foley, Wil Roebroeks, Acheulian and Middle Paleolithic handaxes. the developmental mechanism may J Hum Evol 61:61–74. thus gradually have resulted in a spe- Dennis Sandgathe, John Shea, and 13 White MJ. 1998. On the significance of cific, adaptive, extended phenotype, numerous other colleagues, including Acheulian biface variability in southern Britain. with basic technological behaviors several anonymous reviewers, for their Proc Prehist Soc 64:15–44. becoming innate. assistance with this paper. We also 14 McNabb J. 2005. Hominids and the Early- Middle Pleistocene transition: evolution, culture Another issue that we have left thank various academic audiences for and climate in Africa and Europe. In: Head open is the nature of the Oldowan121 their feedback on presentations in MJ, Gibbard PL, editors. Early-Middle Pleisto- and the recently discovered Lomek- which we have discussed the ideas out- cene transitions: the land-ocean evidence. Lon- 122 don: Geological Society. p 287–304. wian, the industries that preceded lined here. Lastly, we thank the editor 15 Nicoud E. 2013 Le phenome `ne acheuleen en the Acheulean. It is tempting to of Evolutionary Anthropology,John europe occidentale: approche chronologique, argue that if the Acheulean handaxe Fleagle, for his encouragement in the technologie lithique et implications culturelles. Paris: Editions du Comite des travaux histori- was under genetic control, then the face of considerable skepticism on the ques et scientifiques. much simpler tool types of the Oldo- part of certain reviewers, as well as for 16 Richerson PJ, Boyd R. 2005. Not by genes wan and Lomekian were probably his great patience. KV is supported by alone: how culture transformed human evolu- also under genetic control. However, the Netherlands Organization for Sci- tion. Chicago: University of Chicago Press. 17 McNabb J. 2005. More on Acheulean tools. as we explained earlier, McEl- entific Research. MC is supported by 52 Reply to Hodgson. Curr Anthropol 46:648–649. reath’s finding that different mech- the Canada Research Chairs Program, 18 Foley RA. 1987. Hominid species and stone anisms underpin different beliefs in the Social Sciences and Humanities tools assemblages: how are they related? Antiq- his study groups means that there is Research of Canada, the Canada Foun- uity 61:380–392. no reason to expect that all tools in 19 Mithen S. 1999. Imitation and cultural dation for Innovation, the British change: a view from the Stone Age, with spe- the same industry involved the same Columbia Knowledge Development cific reference to the manufacture of handaxes. transmission mechanisms. Needless Fund, and Simon Fraser University. In: Box HO, Gibson KR, editors. Mammalian to say, assuming that three entire social learning: comparative and ecological per- spectives. Cambridge: Cambridge University industries involved the same trans- REFERENCES Press. p 389–399. mission mechanisms is even more 20 Shipton C, Petraglia MD, Paddayya K. 2009. problematic. Given this, we think the 1 Beyene Y, Katoh S, WoldeGabriel G, et al. Inferring aspects of Acheulian sociality and 2013. The characteristics and chronology of the cognition from biface technology. In: Adams applicability of the genetic transmis- earliest Acheulean at Konso, Ethiopia. Proc B, Blades S, editors. Lithic materials and Palae- sion hypothesis should be investi- Natl Acad Sci USA 110:1584–1591. olithic societies. Chichester: Wiley-Blackwell. p gated in relation to pre-Acheulean 2 Deino AL, McBrearty S. 2002. 40Ar/39Ar dat- 219–231. industries, but we are uncertain ing of the Kapthurin Formation, Baringo, 21 Shennan SJ, Steele J. 1999. Cultural learn- Kenya. J Hum Evol 42:185–210. ing in hominids: a behavioural ecological about what will be found. It is possi- 3 Clark Y, Beyene G, WoldeGabriel G, et al. approach. In: Box HO, Gibson KR editors. ble that stone-tool technology went 2003. Stratigraphic, chronological and behav- Mammalian social learning: comparative and 18 Corbey et al. ARTICLE

ecological perspectives. Cambridge: Cambridge 41 Whiten A. 2014. Animal behaviour: incipient editors. Rethinking the human revolution. University Press. p 367–388. tradition in wild chimpanzees. Nature 514:178– Cambridge: McDonald Institute Monographs. 22 Bar-Yosef O. 2006. The known and the 179. 63 Vaesen K. 2012. The cognitive bases of unknown about the Acheulean. In: Goren- 42 Laland KN, Hoppitt W. 2003. Do animals human tool use. Behav Brain Sci 35:203–218. Inbar N, Sharon G, editors. Axe age: Acheulian have culture? Evol Anthropol 12:150–159. 64 d’Errico F, Henshilwood C, Lawson G, et al. tool-making from quarry to discard. London: 43 Laland KN, Janik VM. 2006. The animal cul- 2011. Archaeological evidence for the emer- Equinox. p 479–494. tures debate. Trends Ecol Evol 21:542–547. gence of language, symbolism, and music: an 23 Lycett SJ. 2008. Acheulean variation and 44 Galef BG. 2009. Culture in animals? In: alternative multidisciplinary perspective. selection: does handaxe symmetry fit neutral Laland K, Galef BG, editors. The question of J World Prehist 17:1–70. expectations? J Archaeol Sci 35:2640–2648. animal culture. Cambridge, MA: Harvard Uni- 65 Zilhao J. 2007. The emergence of ornaments 24 Lycett SJ, von Cramon-Taubadel N. 2008. versity Press. p 222–246. and art: an archaeological perspective on the Acheulean variability and hominin dispersals: a 45 Laland K, Kendal JR, Kendal RL. 2009. Ani- origins of “behavioural modernity.” J Archaeol model-bound approach. J Archaeol Sci 35:553– mal culture: problems and solutions. In: Res 54:1–54. 562. Laland K, Galef BG, editors. The question of 66 Villa P, Roebroeks W. 2014. Neandertal 25 Cole JN. 2011. Hominin cognitive and animal culture. Cambridge, MA: Harvard Uni- demise: an archaeological analysis of the mod- behavioural complexity in the Pleistocene: versity Press. p 174–197. ern human superiority complex. PLoS ONE 9: assessment through identity, intentionality and 46 Langergraber KE, Boesch B, Inoue E, et al. e96424. visual display. Unpublished doctoral disserta- 2011. Genetic and “cultural” similarity in wild 67 Henrich J. 2004. Demography and cultural tion, University of Southampton, UK. chimpanzees. Proc R Soc B London 278:408– evolution. How adaptive cultural processes can 26 Mithen S. 1994. Technology and society dur- 416. produce maladaptive losses: the Tasmanian ing the Middle Pleistocene: hominid group size, 47 Lycett S, Collard M, McGrew WC. 2007. case. Am Antiq 69:197–214. social learning and industrial variability. Phylogenetic analyses of behavior support exis- 68 Powell A, Shennan S, Thomas M. 2009. Late Cambr Archaeol J 4:3–32. tence of culture among wild chimpanzees. Proc Pleistocene demography and the appearance of 27 Wynn T, Tierson F. 1990. Regional compari- Natl Acad Sci USA 104:17588–17592. modern human behaviour. Science 324:1298– son of the shapes of later Acheulean handaxes. 48 Lycett SJ, Collard M, McGrew WC. 2009. 1301. Am Anthropol 92:73–84. Cladistic analysis of behavioural variation in 69 Querbes A, Vaesen K, Houkes WN. 2014. 28 Roe D. 2003. An overview, with some wild Pan troglodytes: exploring the chimpanzee Complexity and demographic explanations of thoughts on the study of bifaces. In: Soressi culture hypothesis. J Hum Evol 57:337–349. cumulative culture. PLoS ONE 9:e102543. M, Dibble H, editors. Multiple approaches to 49 Lycett SJ, Collard M, McGrew WC. 2010. 70 Vaesen K. 2012. Cumulative cultural evolu- the study of bifacial technologies. Philadelphia: Are behavioural differences among wild chim- tion and demography. PLoS ONE 7:e40989. University of Pennsylvania Museum of Archae- panzee communities genetic or cultural? An 71 Collard M, Kemery MD, Banks SJ. 2005. ology and Anthropology Press. p 273–285. assessment using tool-use data and phyloge- Causes of toolkit variation among hunter-gath- 29 Wenban-Smith FF. 2006. Handaxe typology netic methods. Am J Phys Anthropol 142:461– erers: a test of four competing hypotheses. Can and Lower Palaeolithic cultural development: 467. J Archaeol 29:1–19. ficrons, cleavers and two giant handaxes from 50 Lycett SJ, Collard M, McGrew WC. 2011. 72 Collard M, Buchanan B, O’Brien MJ, et al. Cuxton. Lithics 25:11–21. Correlations between genetic and behavioural 2013. Risk, mobility, or population size? Driv- 30 Machin AJ. 2009. The role of the individual dissimilarities in wild chimpanzees (Pan troglo- ers of technological diversity among recent agent in Acheulean biface variability: a multi- dytes) do not undermine the case for culture. western North American hunter-gatherers. factorial model. J Soc Archaeol 9:35–58. Proc R Soc B 278:2091–2093. Philos Trans R Soc B 368:20120412. 31 Eerkens JW, Lipo CP. 2005. Cultural trans- 51 Langergraber KE, Boesch C, Inoue E, et al. 73 Collard M, Buchanan B, O’Brien MJ. 2013. mission, copying errors, and the generation of 2011. Genetic and “cultural” similarity in wild Population size as an explanation for patterns variation in material culture and the archaeo- chimpanzees. Proc R Soc B 278:408–416. in the Paleolithic archeological record: More logical record. J Anthropol Archaeol 24:316– 52 McElreath R. 2004. Social learning and the caution is needed. Curr Anthropol 54 S8: S388– 334. maintenance of cultural variation: an evolution- S396. 32 Kempe M, Lycett S, Mesoudi A. 2012. An ary model and data from East Africa. Am 74 Anderson C, Read D. n.d. The evolution of experimental test of the accumulated copying Anthropol 106:308–321. cultural complexity: not by the treadmill alone. error model of cultural mutation for Acheulean 53 Boyd R, Richerson PJ. 1985. Culture and Curr Anthropol. In press. handaxe size. PLoS ONE 7:e48333. the evolutionary process. Chicago: Chicago Uni- 75 Buchanan B, O’Brien MJ, Collard M. n.d. 33 Yamakoshi G, Sugiyama Y. 1995. Pestle- versity Press. Drivers of technological richness in prehistoric pounding behavior of wild chimpanzees at Bos- 54 Boyd R, Richerson PJ. 1996. Why culture is Texas: an archaeological test of the risk and sou, Guinea: a newly observed tool using common, but cultural evolution is rare. Proc Br population size hypotheses. Archaeol Anthropol behavior. Primates 36:489–500. Acad 88:77–93. Sci. In press. 34 McGrew WC. 1992. Chimpanzee material 55 Potts R. 1996. Evolution and climate vari- 76 Schiffels S, Derbin R. 2014. Inferring culture: implications for human evolution. ability. Science 273:922–923. human population size and separation history Cambridge: Cambridge University Press. 56 Potts R, Faith JT. n.d. Alternating high and from multiple genome sequences. Nat Genet 35 Matsuzawa T. 1994. Field experiments on low climate variability: the context of natural 46:919–925. use of stone tools by chimpanzees in the wild. selection and speciation in Plio-Pleistocene 77 Heyes C. 2012. What’s social about social In: Wrangham RW, McGrew WC, de Waal hominin evolution. J Hum Evol. In press. learning? J Comp Psychol 126:193–202. FBD, et al., editors. Chimpanzee cultures. Cam- 57 de Menocal PB. 2011. Climate and human 78 Hill KR, Walker RS, Bozic ˇevic´ M, et al. bridge, MA: Harvard University Press. p 351– evolution. Science 311:540–541. 2011. Co-residence patterns in hunter-gatherer 370. 58 Maslin MA, Shultz S, Trauth MH. 2015. A societies show unique human social structure. 36 Boesch C, Boesch-Achermann H. 2000. The synthesis of the theories and concepts of early Science 331:1286–1289 chimpanzees of the Ta€ı Forest: behavioural human evolution. Philos Trans R Soc B 370: 79 Durkheim E. 1995. The elementary forms of ecology and evolution. Oxford: Oxford Univer- 20140064. religious life. New York: Free Press. sity Press. 59 Mithen S. 1996. The prehistory of the mind: 80 Sosis R. 2000. Religion and intra-group 37 Whiten A, Goodall J, McGrew WC, et al. the cognitive origins of art and science. Lon- cooperation: preliminary results of a compara- 1999. Cultures in chimpanzees. Nature 399: don: Thames and Hudson. tive analysis of Utopian communities. Cross- 682–685. 60 Klein R. 2000. Archeology and the evolution Cult Res 34:70–87. 38 Whiten A, Goodall J, McGrew WC, et al. of human behavior. Evol Anthropol 9:17–36. 81 Irons W. 2001. Religion as a hard-to-fake 2001. Charting cultural variation in chimpan- 61 Mellars P. 2005. The impossible coinci- sign of commitment. In: Nesse R, editor. The zees. Behavior 138:1481–1516. dence: a single-species model for the origins of evolution of commitment. New York: Russell 39 McGrew WC. 2004. The cultured chimpan- modern human behaviour in Europe. Evol Sage Foundation. p 292–309. zee: reflections on cultural primatology. Cam- Anthropol 14:12–27. 82 Wilson DS. 2002. Darwin’s cathedral: evolu- bridge: Cambridge University Press. 62 Wynn T, Coolidge F. 2007. Did a small but tion, religion and the nature of society. Chi- 40 Whiten A. 2011. The scope of culture in significant enhancement in working memory cago: University of Chicago Press. chimpanzees, humans and ancestral apes. capacity power the evolution of modern think- 83 Sosis R, Ruffle BJ. 2003. Religious ritual Philos Trans R Soc B 366:997–1007. ing? In: Mellars P, Boyle K, Bar-Yosef O, et al., and cooperation: testing for a relationship on ARTICLE The Acheulean Handaxe: More Like a Bird’s Song Than a Beatles’ Tune? 19

Israeli religious and secular kibbutzim. Curr theory for vocal learning origin. PLoS ONE 3: 116 Durham WH. 1991. Coevolution: genes, Anthropol 44:713–722. e1768. culture, and human diversity. Stanford, CA: 84 Xygalatas D. 2013. Effects of religious set- 101 Scharff C, Adam I. 2013. Neurogenetics of Stanford University Press. ting on cooperative behaviour: a case study birdsong. Curr Opin Neurobiol 23:29–36. 117 Wheeler M, Clark A. 1999. Genic represen- from Mauritius. Relig Brain Behav 3:91–102. 102 Brenowitz EA, Bleecher MD. 2005. Song tation: reconciling content and causal complex- 85 Schloss JP, Murray MJ. 2011. Evolutionary learning in birds: diversity and plasticity, ity. Br J Philos Sci 50:103–135. accounts of belief in supernatural punishment: opportunities and challenges. Trends Neurosci 118 Wheeler M, Clark A. 2008. Culture, embod- a critical review. Relig Brain Behav 1:46–99. 28:127–132. iment and genes: unravelling the triple helix. 86 Norenzayan A. 2013. Does religion make 103 Kenward B, Rutz C, Wiet AAS, et al. 2006. Philos Trans R Soc B 363:3563–3575. people moral? Behaviour 151:365–384. Development of tool use in New Caledonian 119 Sterelny K. 2004. Review of evolution and 87 Movius HL, Jr. 1948. The Lower Palaeolithic crows: inherited action patterns and social learning: the Baldwin effect reconsidered, cultures of southern and eastern Asia. Trans influences. Anim Behav 72:1329–1343. Weber B, Depew D, editors. Evol Dev 6:295– Am Philos Soc 38:329–420. 104 Borsar A, Ottoni EB. 2005. Preliminary 300. 88 Lycett SJ, Bae CJ. 2010. The Movius Line observations of tool use in captive hyacinth 120 Weber B, Depew D, editors. 2003. Evolu- controversy: the state of the debate. World macaws (Anodorhynchus hyacinthinus). Anim tion and learning: the Baldwin effect reconsid- Archaeol 42:521–544 Cognit 8:48–52. ered. Cambridge, MA: MIT Press. 89 Zhang P, Huang W, Wang W. 2010. Acheu- 105 Tebbich S, Taborsky M, Fessl B, et al. 121 Schick K, Toth NP. 2006. An overview of the lean handaxes from Fengshudao, Bose sites of 2001. Do woodpecker finches acquire tool-use Oldowan industrial complex: the sites and the South China. Quatern Int 223–224:440–443. by social learning? Proc R Soc Lond B 268: nature of their evidence. In: Toth NP, Schick K, 2189–2193. editors. The Oldowan: case studies into the ear- 90 Mesoudi A, Whiten A, Dunbar R. 2006. A 106 Thouless CR, Fanshawe JH, Bertram BCR. liest Stone Age. Bloomington: Stone Age Insti- bias for social information in human cultural tute Press. p 3–42. transmission. Br J Psychol 97:405–423. 1989. Egyptian vultures Neophron percnopterus and ostrich Struthio camelus eggs: the origins 122 Harmand S, Lewis JE, Feibel CS, et al. 91 Lopez-Maury L, Marguerat S, Bahler€ J. of stone-throwing behaviour. Ibis 131:9–15. 2015. 3.3-million-year-old stone tools from 2008. Tuning gene expression to changing envi- 107 Hansell M. 2000. Bird nests and construc- Lomekwi 3, West Turkana, Kenya. Nature 521: ronments: from rapid responses to evolutionary tion behaviour. Cambridge: Cambridge Univer- 310–315. adaptation. Nat Rev Genet 9:583–593. sity Press. 123 Roebroeks W, Corbey R. 2001. Biases and 92 Sanderson MJ, Hufford L, editors. 1996. 108 Hansell M. 2007. Built by animals: the nat- double standards in palaeoanthropology. In: Homoplasy: the recurrence of similarity in evo- ural history of animal architecture. Oxford: Corbey R, Roebroeks W, editors. Studying lution. New York: Academic Press. Oxford University Press. human origins: disciplinary history and episte- 93 Lockwood CA, Fleagle JG. 1999. The recog- mology. Amsterdam: Amsterdam University 109 Madden JR. 2008. Do bowerbirds exhibit nition and evaluation of homoplasy in primate Press. p 67–76. cultures? Anim Cognit 11:1–12. and human evolution. Am J Phys Anthropol 124 Roberts MB, Parfitt SA, Pope, MI, et al. 110 Penn D, Povinelli DJ. 2007. Causal cogni- 110:189–232. 1997. Excavation of the Lower Palaeolithic site tion in human and nonhuman animals: a com- 94 Collard M, Wood BA. 2001. Homoplasy and at Amey’s Eartham Pit, Boxgrove, West Sussex. parative, critical review. Ann Rev Psychol 58: the early hominid masticatory system: infer- Proc Prehist Soc 63:303–358. 97–118. ences from analyses of living hominoids and 125 Sharon, G. 2011. New Acheulian locality 111 Gowlett JAJ. 2006. The elements of design papionins. J Hum Evol 41:167–194. north of Gesher Benot Ya’aqov: contribution for form in Acheulian bifaces: modes, modalities, Iacoboni M, Dapretto M. 2006. The mirror the study of the Levantine Acheulian. In: J.-M. 95 rules and language. In: Goren-Inbar N, Sharon Le Tensorer, R. Jagher, M. Otte, editors. The neuron system and the consequences of its dys- G, editors. Axe age: Acheulian tool-making from Lower and Middle Palaeolithic in the Middle function. Nat Rev Neurosci 7:942–951. quarry to discard. London: Equinox. p 203–221. East and neighboring regions. Lie`ge: Centre de 96 Keysers C, Gazzola V. 2006. Towards a uni- 112 Soressi M. 2002. Le Mousterien de Tradi- fying neural theory of social cognition. Prog recherches archeologiques de l’Universitede tion Acheuleenne du sud-ouest de la France. Lie`ge. p 25–33. Brain Res 156:379–401. Bordeaux: Universite Bordeaux I. 97 Del Giudice M, Manera V, Keysers C. 2008. 126 Vaufrey R. 1955. Prehistoire de l’Afrique, 113 Joris€ O. 2003. Zur chronostratigraphischen vol I. Paris: Masson. Programmed to learn? The ontogeny of mirror Stellung der spatmittelpal€ aolithischen€ Keilmes- 127 Kuman K. 1994. The archaeology of Sterk- neurons. Dev Sci 12:350–363. sergruppen. Versuch einer kulturgeographi- fontein — past and present. J Hum Evol 27: 98 Stout D, Chaminade T. 2009. Making tools schen Abgrenzung einer mittelpalaolithischen€ 471–495. and making sense: complex, intentional behav- Formengruppe und ihr europaischer€ Kontext. iour in human evolution. Cambr Archaeol J 19: Berlin Rom.-German.€ Kommission 84:49–153. 128 Leakey M. 1971. Olduvai Gorge, vol III: excavations in Beds I and II, 1960–1963. Cam- 85–96. 114 Ruebens K. 2013. Regional behaviour bridge: Cambridge University Press. 99 Stout D, Toth N, Schick K, et al. 2008. Neu- among late Neanderthal groups in Western ral correlates of Early Stone Age toolmaking: Europe: a comparative assessment of late Mid- 129 Paddayya K, Jhaldiyal R, Petraglia MD. technology, language and cognition in human dle Palaeolithic bifacial tool variability. J Hum 2006. The Acheulian quarry at Isampur, Lower evolution. Philos Trans R Soc B1499:1939– Evol 65:341–62 Deccan, India. In: Goren-Inbar N, Sharon G, editors. Axe age: Acheulian toolmaking from 1949. 115 Cavalli-Sforza LL, Feldman MW. 1976. quarry to discard. London: Equinox. p 45–74. 100 Feenders G, Liedvogel M, Rivas M, et al. Cultural transmission and evolution: a quanti- 2008. Molecular mapping of movement- tative approach. Princeton, NJ: Princeton Uni- associated areas in the avian brain: a motor versity Press. VC 2016 Wiley Periodicals, Inc.