Technological sophistication of the earliest toolmakers
One study reported at the meetings brought to mind a growing literature on the sophistication of Pliocene archaeological assemblages, grouped as "Oldowan" or "pre-Oldowan." Most Oldowan tools lack a standardized form, and the assemblages are dominated by flakes, core tools with one sharp edge or point, and unmodified manuports or hammerstones. The lack of standardized tools (such as handaxes) lead easily to the conclusion that Oldowan technical assemblages are simple, so that anyone could make one who could chip rock.
But a number of recent sources illustrate that this perception isn't accurate. The Oldowan encompasses a variation of manufacturing complexity and other elements, but cannot be said to be merely rocks that have been carried or bashed together. The most interesting studies have examined the issue of raw material exploitation and the actual processes used to reduce stone into useful tools.
David Braun gave a presentation that relates to this topic at the 2005 Paleoanthropology Society meetings. In an analysis of the lithic remains at the Kanjera South site (Kenya), the group found that the intensity of the reduction sequence varied with source material.
Kanjera South, dating to around 2.2 million years, preserves one of the largest early records of material culture. According to Plummer et al. (2001:810):
Hominins at Kanjera South utilized a wider variety of lithic raw materials than found at most Oldowan sites, some of which (chert, quartz, quartzite) must have been transported from outside the immediate vicinity of the deposits since they are not present in the local clast population.
In the study by Braun et al. (2005), fine-grained quartzite, which had a nearest source outcrop a long distance away from the site, was reduced more intensively than locally available rock. This reduction involved more flakes taken per core, and occasional retouch. They hypothesized that this difference was due to the hardness of the quartzite and its ability to hold an edge, which makes tools made on quartzite potentially more useful. The hominids apparently recognized these qualities in the raw materials and exploited the quartzite. We can probably infer that the transport distance was a result of this recognition as well.
Stout and colleagues (2005) report on the use of raw materials at Gona, which is at present the earliest site at which stone tools are found, dating to around 2.6 million years. This study sampled unmodified stone taken from conglomerates that represent natural accumulations. If stone were taken merely from local sources without selectivity, the modified stone assemblages from the site would be expected to match these conglomerates. As described by Stout et al. (2005:366-367), the past understanding of Oldowan toolmakers has suggested that they may not have exhibited a clear understanding of the technical qualities of their raw materials:
At Koobi Fora, the high frequency of basalt in local conglomerates (i.e. ancient stream channels) is closely mirrored by its predominance in the artifact assemblages (Toth, 1985; Schick, 1987; Isaac et al., 1997). Although quartz, chert and glassy volcanics are "reasonably easily available" in the conglomerates, these material types do not appear to have been specifically selected for (Isaac et al., 1997:268).
This paper gives an excellent review of the history of raw material choice. In brief, the literature review indicates that early toolmakers at Koobi Fora may have exercised some degree of selectivity by avoiding certain kinds of stone (vesicular lavas and weathered cobbles). By Olduvai Bed II (after 1.7 million years ago), some toolmakers were clearly going out of their way to acquire certain stone materials:
The increased use of chert at Olduvai around 1.65 Myr was clearly occasioned by the temporary exposure of rich sources of this material by the retreating waters of the paleo-lake (Hay 1976). Hominid toolmakers at this time readily appreciated the superior flaking properties of chert, leading to the formation of the earliest known special-purpose quarry site at MNK CFS (Stiles et al. 1974; Stiles 1998) (Stout et al. 2005:367).
So when did this technical knowledge really originate? Was it a skill that developed gradually over the million years between Gona and Olduvai Bed II?
Stout and colleagues find that the materials used for tools at Gona do not match the natural cobbles that were avaiable in a number of respects. In essence, the hominids ignored many kinds of stone that had fracture patterns or other properties that made them less suitable as raw materials. They (2005:368) state the conclusion most clearly at the end of their introduction:
The strong pattern of raw material selection seen at Gona demonstrates that low levels of selectivity are not universal in the Oldowan. The fact that the assemblages from Gona are the oldest known in the world also argues against an overarching temporal trend toward increasing selectivity within the Oldowan.
The flaking skill of early hominids is the subject of a paper by Roche and colleagues (1999) in Science. At the two localities of Lokalalei 1 and 2C, both dating to 2.34 million years ago on the west shore of Lake Turkana, this study accomplished refits of stone flakes and cores to directly inspect the reduction sequence. They found that this sequence was quite complicated in many of the instances of manufacture at the two localities:
At LA2C, the dominant reduction sequence consists of these more reduced cores. Unidirectional or multidirectional removals are flaked on a single debitage surface, from natural or prepared platforms. This knapping system allows the production of a long series of removals without changing the volumetric structure of the core. The repeated application by the knappers of the same technical principles to a whole series of cores, and during the reduction of each core, indicates an elaborate debitage scheme, implying motor precision and coordination. These principles include an appreciation of the quality of the collected raw materials, a judicious exploitation of the natural morphology of the blocks and the maintenance of adequate flaking angles during the entire debitage sequence. These show that the notion of production was already assimilated by a group of hominids in this particular area. This notion is integrated within a real debitage strategy, here well-mastered and unprecedented for this time period (Roche et al. 1999:59).
Upon this description and analysis of the debitage, the study bases the following conclusion:
Overall simplicity and similarities between assemblages are the two main arguments recently put forward to substantiate a technological stasis hypothesis between the 2.6 and 1.6 Myr time periods, and to merge the related assemblages into a single vast 'Oldowan' technocomplex. The stasis hypothesis cannot hold out against the detailed technological analysis of the LA2C lithic assemblage. There can be no doubt about the elaborate character of the LA2C debitage schemes, which are far more sophisticated than at any other Pliocene site (Roche et al. 1999:59).
In total, these studies attest to the technological expertise of some early toolmakers. It is perhaps important to remember that the archaeological record should be expected to reflect an average technical competence much less than that possible (or possibly even typical) of early hominids. For example, every instance of stone flaking by an expert toolmaker certainly was preceded by novice flaking by the same individual earlier in his life. Considering the early mortality evidenced by australopithecine (and early human) remains, many potential toolmakers must have died before reaching proficiency. And toolmaking expertise was probably not uniformly distributed among social groups, so that some individuals had the opportunity to learn by observation some relatively complex reduction sequences and material properties, while others would have had to acquire the same knowledge by long trial and error, if at all.
In this framework, later changes in the archaeological record may reflect social and life history changes as much as ecological changes or brain evolution. Toolmaking intensifies later in the Oldowan, raw material selectivity increases, and some standard forms and more technically sophisticated artifacts, such as bifaces, routinely occur. The increased mental complexity of Pleistocene hominids may be a cause, but the behavioral changes have no obvious anatomical, life history, or other correlate. The apparent sophistication of the earliest toolmakers argue against the hypothesis that the later Oldowan represents a fundamentally more intelligent hominid species, with specialized adaptations to tool manufacture lacking in earlier hominids. Instead, I would suggest that later Oldowan hominids had on average a greater retention of cultural knowledge as a function of more stable groups and a longer life history than australopithecines. This shift might or might not have been accompanied by an increase in the capacity to learn, as might be evidenced by a greater length of the juvenile growth period.
From a conceptual perspective, there is necessarily a real limit on the standardization that is possible within any "technocomplex," and particularly one that spans a million years or longer. The kind of information that is ultimately transmissible (either horizontally or vertically) among hominids over this extreme time period is of the most rarified form imaginable. A model of early archaeological assemblages that incorporated the probable manner and form of information transfer among groups (along with their global demography) would predict both a global lack of complexity and a great extent of variability within that simple context. The evidence of technical skills therefore may inform directly about the demography of groups and their relationships with each other. The extent to which such variability may be consistent with cultural diffentiation among groups in other primate species (such as chimpanzees) I take to be a testable hypothesis.
References:
Most references within the featured articles below:
Plummer T, Ferraro J, Ditchfield P, Bishop L, Potts R. 2001. Late Pliocene Oldowan excavations at Kanjera South, Kenya. Antiquity 75:809-810.
Roche H, Delagnes A, Burgal JP, Feibel C, Kibunjia M, Mourre V, Texler PJ. 1999. Early hominid stone tool production and technical skill 2.34 Myr ago in West Turkana, Kenya. Science 399:57-60.
Stout D, Quade J, Semaw S, Rogers MJ, Levin NE. 2005. Raw material selectivity of the earliest stone toolmakers at Gona, Afar, Ethiopia. J Hum Evol 48:365-380.
Whallon's population model
Whallon R. 1989. Elements of cultural change in the later Paleolithic. In: Mellars P, Stringer CB, editors, The human revolution: Behavioural and biological perspectives on the origins of modern humans. Edinburgh: Edinburgh University Press. p 433-454.
I was really privileged to have Bob Whallon's course when I was at Michigan. It certainly taught me almost everything I know about population structure in foraging populations, and if not for that knowledge, I would probably not have returned to consider genetic variation from the perspective of population structure. Now I am pursuing topics much closer to Whallon's own interests, in the relation of information theory to group structure in ancient humans. This article, from the Human Revolution volume, is the best statement of Whallon's thinking on these issues.
He introduces the paper with one of the central issues of prehistory: why do so many things appear to happen essentially simultaneously and independently in different regions of the world during the Upper Paleolithic? The case in point is the migration of people to Australia and the New World, and Whallon's explanation is organizational and cognitive (433):
These explanations postulate not only the emergence of new socio-cultural structures as others also have proposed (Gamble 1983), but that these structures, in turn, would have required the development at this time of even more fundamental human capacities for conceptualization and communication.
Whallon builds his argument from a number of assumptions, and these are worth enumerating:
- Human population densities are ultimately resource-limited. Beyond some minimum population density, human populations cannot maintain themselves in a geographic region, because of the impossibility of maintaining breeding relations. This point relies heavily upon the early work of Martin Wobst, who attempted to determine the minimum sizes possible for foraging groups under the condition that mates must be available for maturing individuals (Another review available here).
- "[S]cales of of geographical movement and contact among hunter-gatherer groups" are themselves conditioned upon the kinds of foraging strategies that are within the cognitive range of a human group. Some environments fluctuate wildly, and may necessitate more in the level of movement and alternation of resource base than would have been possible for humans at a given adaptive level.
- Conditions (1) and (2) together imply that some large territories will be simply uninhabitable, because of the constraints on human populations in ecological adaptability and the numbers necessary to maintain occupation.
- Some constraints may be overcome by the application of suitable communication and movement strategies. If appropriate strategies became available, they would allow the exploitation of formerly uninhabitable regions.
- Low-density populations face informational limits. When extensive movement and a wide geographic range for groups is necessary, it becomes less possible for individuals to have "direct knowledge and immediate, face-to-face contact" with other individuals and groups, including potential mates, trading partners, and competitors.
- An essential feature of modern human language is "displacement." This skill is necessary to refer to objects and relationships not immediately present, and thereby forms a substitute to the direct, face-to-face communication that would otherwise be the only way to convey such concepts.
- Another essential element of human social life is the ability to synchronize social expectations among groups, at least to a degree necessary to readily facilitate the movement and establishment of individuals into new groups. Whallon contrasts the human situation with the frequently troubled and violent integration of dispersing primates into new groups. In his view, humans have alliance networks to facilitate such social integration, which are themselves dependent on the displacement function of language for the communication of expectations and abstract categories accompanying such human relations. Whallon observes that kinship categores are among this class of abstract relations that exist in all human groups.
- Decision-making ability is maximized in human groups when the number of decision-makers is around six (Johnson 1978; Reynolds 1984). Beyond this size there is strong pressure for hierarchicalization of decision-making.
Whallon sees several aspects of human societies as outgrowths of this change in communication ability. These include kinship, collective foraging strategies, division of labor, ability to plan for the future, and oral history. This last may be the most important, because Whallon sees it as an adaptation to adjust to temporal fluctuations in resource access. In brief, if people can remember the last time something bad happened (such as a drought) and can remember what they did to adjust, they can communicate that information to their children or grandchildren for use next time similar events happen.
Moving from a situation in which knowledge of the past is restricted to individual memories to the sharing of such memories through an ability to refer to them and talk about them creates something new: a group memory. Group memory provides entirely new opportunities, and results in the devleopment of a number of cultural mechanisms for information storage, retrieval, and use, whose roles within cultural systems are complex but highly important (cf. Mine 1986). Without going into the details of these, we only will consider briefly here the fact that the fundamental effect of any and all of these mechanisms, beginning with simple verbally-shared individual memories, the basic and possibly earliest form of group memory, is that they increase the time depth of information available to the members of any group significantly beyond the life-span of an individual.
The most obvious effect of this expansion of the time depth of information available to a hunter-gatherer group is that it allows adaptation to environmental fluctuations with a periodicity greater than the average individual life span (444).
Whallon gives an estimate of this time span, figuring it to be roughly equal to three to five generations, or 60 to 125 years. He places oral history, myth, and other means of recalling information over long times as elements of this information storage system.
He also puts together two opposing ideas from the information standpoint that may constrain the size of hunter-gatherer groups.
It is perhaps in this light that the universally observed number of some 25-30 members (6 nuclear families) in recent hunter-gatherer minimal bands may be seen to have an organizational basis. Both the increasing quality of decisions with the involvement of more decision-makers (Johnson 1978) and the exponentially decreasing risk of group extinction (Wobst 1974: 171-172) would confer advantages to larger groups up to the level of six family units, beyond which it becomes increasingly difficult to maintain both effective decision-making and an egalitarian organization.
Egalitarian organization is a fundamental aspect of hunter-gather organization that is probably impossible to change, considering the lack of resource hoarding or other means of maintaining power inequities. Larger groups would almost certainly fission if they could do so, rather than allow voluntary resource inequities to perpetuate themselves. Thus, the optimum number of decision-makers might be assumed to be a mental adaptation to the kinds of groups that were typical for extrinsic reasons. Under this assumption, the system would continue indefinitely because of the reinforcement from both mental and ecological realities.
One of the most interesting emerging ideas here I would call the "information economy." Whallon's model implicitly postulates two stages, with a threshold between them. In the earlier stage information is highly valuable, but it can be obtained only through direct perception or "immediate, face-to-face communication." Whallon allows that this earlier stage may include symbolic communication, particularly of the sort encompassed by Bickerton's (1990) concept of a protolanguage.
It is in the later stage that information can itself function as a commodity.
Thus, it would seem that two of the important blocks to human expansion at any given level of technology into environments of low resource density and predictability are defined by communicational and social abilities: the ability to know about mates and resources within the physically attainable region, and the ability to assure access to these once known....On the other hand, the development of the communicational abilities necessary for: (1) the adequate transfer of information, and (2) the creation of the social networks necessary for such information flow, as well as for assurance of access to mates and resources over an adequately large area and sufficient number of local groups, both seem to repose on the emergence of a roughly equivalent level of language competence (440).
If information can be transmuted between these two symbolic categories--knowing about things and assuring access to things--then information itself may bear a value extrinsic to its own initial generation. This economy is based on a currency whose value is constantly uncertain, since the intrinsic value of any single piece of information may be unknown to sender, recipient, or both. This dynamic establishes a dynamic in which cooperation is much more beneficial than defection, since the value of any single piece of information may be unknown to one who holds it. Only by pooling information maximally will the possible benefits of information be maximized.
On the other hand, a very intelligent agent may be able to judge the value of information more accurately. And it may become possible to withhold information strategically, exchanging it only for other information (remember that this includes assurances of resource access) of great value. And one might be able to manipulate the symbolic system to generate spurious information that is nonetheless believed to have great value. The attribution of malevolent events to "witches," the propitiation of certain gods, and other concepts of natural forces no doubt fall within this category of information phenomena. Playing the information economy in this way would lead to a selection pressure for manipulation of the symbolic space, as well as for accurate detection of such manipulation or cheating.
Potential problems
There are a few weak points in Whallon's scenario. For example (436):
Inability to obtain adequate information, either in terms of reliability and accuracy or of the area covered, on the existence and location of needed mating partners or subsistence resources, subjects individuals and groups to risks of failure to reproduce at a sufficient rate or of nutritional stress and eventually starvation. As outlined above, such risks and the penalties for failure grow with decreasing resource density and predictability.
This is intuitively appealing, but ultimately relies on a fallacy about population density. Sure, low-density human populations were faced with resource stress. But high-density populations were almost certainly under just as much resource stress. Compared to other hominoids, humans have a high intrinsic rate of population increase, and we can expect that they were near the maximum possible population density in most of their range. This means that the competition for resources was greater in higher-density regions, not less. And it almost certainly means that whether peripheral populations in marginal regions lived or died had little relevance to the largest proportion of ancient humans. It may be true that habitation of low-resource-density regions required advances in communication and cognition. But when humans lived in both low-density and high-density regions, they probably faced equal or greater resource stress in high-density regions.
It is natural to focus on the expanding populations in low-density environments, because these new populations are one of the major gross differences in the archaeological record of the latest Pleistocene when compared with earlier time periods. But the most important changes likely occurred in the heart of the long-occupied human range. These changes are more or less invisible, since the important ones involved not artifact types but instead organizational principles. However, we can infer that they happened based on Whallon's argument or another like it.
Another issue is the degree to which hunter-gatherer populations track their environments.
One result of these shifts, universally observable today among hunter-gatherers and the source of frequent comment and speculation, is that human population densities would appear more and more of the time to be below the supposed 'carrying capacity' of their environment. The resource 'lows' to which the existence of a shared, group memory allows a population to adapt would be less frequent, with the consequence that periods of resource stress due to fluctuations that reach or exceed a group's abilities to respond effectively would occur at wider intervals. The phenomenon of adaptive responses being tuned to longer-term, larger resource fluctuations, so that field studies and observation (including ethnographic expeditions) normally find populations under non-stressed conditions in which many important adaptive responses are latent and cannot be seen in operation, is well known also in the ecological study of many species (445).
This passage appears to be making two distinct arguments. The first is that human populations are adapted to long-term resource fluctuations, so that they typically maintain themselves at lower densities than can usually be maintained. Their "adaptations" prepare them for future resource stress, in other words, by limiting their population growth. This argument is simply nonsense from an evolutionary perspective; there is no advantage to anyone's genes in limiting their reproduction in anticipation of population stress far in the future.
The second argument is more subtle. If humans have successfully adapted to low-resource times by pooling information from previous times with similar ecological stress, then they may be able to maintain a larger population during such times and recover more quickly from them. To the extent that humans are self-limited in resource availability, these information adaptations will tend to smooth resource access over time, minimizing fluctuations (at least with respect to their effect on human resources, if not on the absolute productivity of ecosystems or similar measures). This would not cause humans to be below the carrying capacity, however; it would cause humans to more nearly match the carrying capacity over more of the time.
I think the "well below carrying capacity" assumption is basically false. The human population has likely been at or close to its carrying capacity until very recent technological changes. Every evidence from history suggests that human populations grow quickly in the presence of enough resources, and such growth rates expressed across tens of thousands of years could not help but fill areas to their maximum population density. The idea that there is a pool of unused resources depends on false logic about what might be accomplished by a larger population, ignoring the consequences of higher exploitation over longer time periods (and thus the feedback between human populations and their environments) as well as the energetic costs of more intensive foraging. If humans expend 3000 calories or more a day to forage as effective hunter-gatherers, there will be diminishing returns to longer and more intensive foraging, in energetic, cultural and social terms.
The idea that six decision-makers optimizes decision effectiveness is taken from empirical research. But many questions are left unanswered concerning this value. If larger groups were adaptive (because they had significantly longer half-lifes, for example) then why couldn't humans adapt to be better decision makers with eight or ten individuals, instead of six? Is there a fundamental organizational contraint related to time, or argumentation? Or is this a purely mental limit of humans, and some computer might work well with a hundred decision makers or more? Is physical force sitting at the root of this, with some minimum number being necessary to avoid the possibilty of lethal coalitions?
References:
Johnson, G. A. 1978. Information sources and the development of decision-making organizations. In C. L. Redman et al. (eds) Social Archeology: Beyond Subsistence and Dating. New York: Academic Press: 87-112.
Mine, L. 1986. Scarcity and survival: the role of oral tradition in mediating subsistence crises. Journal of Anthropological Archaeology 5: 39-113.
Reynolds, R. G. 1984. A computational model of hierarchical decision systems. Journal of Anthropological Archaeology 3: 159-189.
Wobst's population model
Reference:
Wobst, H. Martin. 1974. Boundary conditions for Paleolithic social systems: a simulation approach. American Antiquity 39(2):147-178. JSTOR
To me, this paper is a landmark. The reason why is that it is a serious attempt to make something out of nearly nothing. The nearly nothing is precisely what we know about the social structure of hunter-gatherers. Wobst has only the barest of basics to begin with: exogamous bands of around 25 individuals, organized into larger groups (here called maximum bands) that are territorial and endogamous. There is little of substance to add here, beyond the size of the maximal band and some idea of its geographic extent, but in a sense these details are reserved for cross-checking the end result of the simulation (although not literally so). The something that Wobst constructs is not a general model for Paleolithic social structure--this model falls directly from his assumptions, and its validity is no greater or less than that of the assumptions themselves. The real accomplishment of the study is the derivation of several other consequences of the assumptions that are not in the least bit obvious corollaries.
Thus, it was surprising to me in a quick check of Google Scholar that the paper does not have more of a following. This is an imperfect comparison, since it focuses only on publications available to the service (i.e. recent and web-accessible), but I think it reflects a general perception about the paper. Most of the people I know are well aware of the paper and its arguments, but few see any practical applicability. In this context, Wobst's apparent concern for contextualization is well-founded. It is all well and good to formulate a general model of what Paleolithic social structure may have been like, but this in effect is a global hypothesis for which none but extremely local tests are really possible.
So who cites the paper? Fundamentally, if you need a model of Paleolithic social structure, this is the study to start with. The fact is that beyond Wobst's references there is very little to go on in terms of concrete ethnographic comparisons. There are critiques of the numerical values taken as Wobst's assumptions, but there is certainly nothing much better to work with. And the logical arguments behind some of the derivations may be quibbled with, but there is no fundamental flaw in them that would require taking up an alternative logic. A global model of population structure is a basic prerequisite for genetic studies of ancient demography, for hypotheses about the ecological relationships of ancient groups, and for studies of the migration and interaction of ancient peoples. They underlie much of evolutionary psychology, with its focus on the social and cultural interactions that may have served as global standards (or rough averages) during human evolution.
Even the preamble of the paper is interesting--it is as concrete a demonstration as one might find for the reality of a "Michigan" school of anthropology. Here we can see Wobst arguing for archaeology as a mode of investigating "cultural systems" and as a key to the development of "general cultural theory" (148). It resonates for the biological anthropologist because at the very same time paleoanthropologists were discussing the power of culture as an active element in shaping human evolution, including the assertion that culture was the "adaptive niche" to which humans were more or less adapted. Ideas like these were by no means limited to Michigan at the time, but in a department with the legacy of Leslie White they had a status as serious statements of cultural theory that they have never again enjoyed. These attempts by archaeologists and biological anthropologists were not the mainstream of cultural theory, which was already in the grips of Derrida, or at least Clifford Geertz. Indeed it is somewhat ironic to read Wobst discussing how Paleolithic studies had moved toward a steadily closer alliance with social studies over the preceding century, when at that very moment the social sciences were moving toward a rejection of most of the ideas Wobst was embracing.
Nevertheless, this preamble does not really serve as a foundation for the argument to follow; rather, it serves to contextualize it. The underlying concern that Wobst takes time to address is whether general theoretical statements about the behavior of past groups may be tested using the shovel-in-the-ground data from real archaeological sites. His position here is that the data increasingly was being used for testing hypotheses of cultural or social content, and so there is no logical barrier to further work along these lines. Wobst also reminds us that theory in archaeology precedes the recovery of better data--instead of the other way around--because everyone would remain satisfied with the previous level of data collection if there were no logical reasons to work more intensively (150).
Minimum and maximum bands
Wobst takes these terms from Julian Steward (1969). They are often treated as equivalent to Joseph Birdsell's categories of band and "tribe." Moreso than for most human social groups, and more akin to social groups in other mammal species, these groups have reality only to the extent that structured behavior reinforces them. In hunter-gatherer groups, the flexibility of the social units is such that neither type of band is static over its lifespan, or even over the lifespan of individual humans. These groups are dynamic in membership, in territory, and in their interactions with other groups.
Significantly, both kinds of bands delineate social networks that are larger than a nuclear family or comparable kin group. This means that the forces that maintain these groups formally occur at a different level from the interactions typical of close kin. For the minimal band, the individuals may be related by more distant kinship or affinal ties, and Wobst notes the kinds of factors that cause minimal bands to persist including, "cultural practises of cooperation among its members, division of labor according to age and sex, and mutual food-sharing" (152). The size of the minimal band is contrained by ecological facts: larger groups are desirable to reduce risk (predation risk, risk of hunting failure, risk of incursion by other groups) while smaller groups are desirable to minimize competition for food and other resources.
Wobst introduces the maximum band as
a marriage network which guarantees the biological survival of its members, since the members of a minimum band have to rely on a larger number of persons than their own membership in order to provide a member with a mate upon reaching maturity (152).
It is the exchange of mates among minimum bands that enables the social integration of the maximum band, to the extent it exists. This includes details like shared cultural rules and language, as well as direct negotiations among members of different minimum bands for access to resources like exotic materials and mates themselves. This integration occurs along religious and ideological lines that differ among different hunter-gatherer cultures.
Given these constraints on maximum bands, Wobst predicts two boundary conditions. The minimal equilibrium size is the size of a maximum band that "will consistently guarantee the presence of a suitable mate for a group member auon reaching maturity" (154). This size depends on the marriage rules of the society--for example, prohibitions upon mating with distantly related kin will decrease the number of suitable mates for an individual and thus require a larger minimal equilibrium size to maintain the same mate access. The second boundary condition is the maximal equilibrium size which is "the maximal number of people which can be consistently integrated by the cultural mechanisms of a given cultural system and which is consistently required for the successful operation of such a cultural system" (154). This latter condition is fairly nondeterminate, but the former is a potentially important constraint: if the "cultural system" cannot integrate a given number of people successfully, then some number of them will undoubtedly fission to create a new group.
This is the stream of thought that currently underlies Robin Dunbar's work on social group size in primates; namely that the maintenance of social groups of a given size requires an investment of time on the part of each individual, and without the innovation of new methods of social control this investment ultimately must limit social group size. Whether human groups face this barrier at around the maximal equilibrium size of hunter-gatherer societies is undetermined, but this is the speculation of Dunbar and others. It is to some extent an unfounded speculation because it rests on the assumption that in the absence of such a constraint on group size, group sizes would be larger than they are. This assumption is unfounded, since it implies that larger groups are better in some way that has not been demonstrated. For example, we might as easily believe that the constraints on group size were not due to time constraints on interaction but instead due to ecological reasons. Another option is that group size is influenced by a balance between within-group competition (for political or economic dominance) and between-group competition (for resources or territory), in just the same way that minimum band size is generally thought to be determined. Under that scenario, there is no special social dynamic occurring within the maximum band, and the emergent characteristics at the maximum band level, such as language dialect, religious ideology, and oral history, are themselves derived from the quasi-historical reality of the group rather than any active role of such groups in normalizing human cultural development. At that, Wobst himself says "It is doubtful, however, whether the maximal equilibrium size has any bearing on the cultural system of Pleistocene hunters and gatherers" (154).
Wobst argues for the primacy of the minimal equilibrium size for Paleolithic groups:
What's up with the hexagons?
The most idiosyncratic of Wobst's assumptions is the formulation of a hexagonal abstraction for the geographic arrangement of groups. The hexagonal model is not problematic: it represents the most efficient packing arrangement for equal-sized units upon a two-dimensional surface. For this, the assumption of hexagons is not strictly necessary except to eliminate the "unaccounted voids" (153) that circular territories leave between neighboring groups. To support the idea of a hexagonal arrangement Wobst provides a brief survey of the number of territorial boundaries for bands of different ethnographic hunter-gatherers, finding an average between 5 and 6--thereby supporting the 6-plex arrangment of hexagons in his model. But this may be a fairly general characteristic that reflects any spatial network, regardless of territory shape: consider that there are 11 contiguous U.S. states west of the Mississippi River that lack an ocean shore or national border, and these states have an average of 5.8 neighbors, despite the fact that most of them obviously approximate rectangles in shape.
Less obvious assumptions
Here are several of the less obvious assumptions used by Wobst but not included in the master list of assumptions. These are foci of interest because they illustrate much about the intuitive model of small-scale societies typically deployed by paleoanthropologists.
- "[E]cological restraints on population agglomerations of hunters and gatherers" (154).
- "[T]he lack of intraband specialization except by sex and age" (154).
- "[I]f a local group leaves the niches it is accustomed to exploit, its chances of success in hunting and gathering decline while its chances of failure due to unpredictable events increase. Thus, long-distance moves would tend to lower the population density and introduce an element of instability into an interregional social network" (153). I wonder if this assumption can be restated as a general rule applying to any information-dependent foraging society. Clearly, moving causes risk and the net effect of risk, averaged over many groups, must be to decrease the population size. We may ask whether the risk of not moving balances the risk of moving at some point, and thereby drive a global model of response to local environmental deterioration, as faced by the Neandertals, perhaps. On the other hand, humans probably never had the choice of large-scale dispersal, as argued by Wobst in the next few paragraphs, because most territories suitable to human occupation were never empty after their initial habitation.
One problem with the paper is its assumption that "Paleolithic" equals band-sized social organization. At some point in the past, it is probably possible to assert that large social units were simply not possible. But this time certainly did not encompass all of the Paleolithic, since Upper Paleolithic peoples to a greater or lesser extent were capable of agglomerating into relatively large village-sized social units for long periods of time, possibly using specialized technological and cultural solutions based on local food and resource abundances such as salmon runs or animal migration choke-points. Large social groups would have enabled social complexity beyond that contemplated here, without requiring extractive methods outside the scope of "hunting and gathering." They may well have been limited to regional importance or importance over limited timescales, in the end constraining their possibilities for growth and increasing sophistication, but nevertheless imply very different things for social controls and "boundary conditions" than envisioned in the band-level model.
This raises the question of whether there was any time that the band-level organization model had real adaptive importance for humans. Even if the majority of humans around the world were organized in this maximum band-minimum band social structure, nevertheless we might imagine that selection was most powerful not in this stable context but in the unstable social contexts accompanying local or temporally constrained large populations. It was in these places were rapid population growth, and subsequent dispersal to more marginal areas, was possible. It was therefore in such small, high-density populations that we might expect to find the most active adaptation toward the mental and social characteristics of recent human societies.
Mexican footprint media
The BBC reports on a BBC program covering Silvia Gonzalez and the Cerro Toluquilla "footprints". (I have a roundup of stories from the initial announcement. )
The new BBC article gives a good account of how they were found and the context they are in. The big problem is the dating:
More controversial still are the dates. Colleagues of Silvia Gonzalez at Oxford University used a technique called optically stimulated luminescence (OSL) that records the last time rocks were exposed to sunlight or heat.
That gave a variety of dates from the overlying sediments, but when applied to small fragments of what looked like brick or burnt clay within the volcanic ash, it produced a date of about 40,000 years.
That initially shocked Dr Gonzalez as it implied by far the earliest evidence of humans in the Americas. But it fitted in with dates of up to 38,000 years based on carbon 14 in shells in the sediments above.
The dispute right now boils down to a "last gasp" by Gonzalez. Paul Renne's group found that the ash layer dates to 1.3 million years by Ar-Ar. But there is always a chance that the layer was actually redeposited by water from an earlier ashfall. At 1.3 million years old, the ash layer ought to have reversed polarity, since the last period of normal polarity began 780,000 years ago. When Renne's team checked the polarity of the ash, they found it does have reversed polarity. So the evidence is very strong that the ash layer is 1.3 million years old.
Here's Gonzalez' argument:
Silvia Gonzalez' view? "We know that there are short-term 'excursions' of the magnetic field, and one of those happened 40,000 years ago, very interestingly."
Why is this special pleading? Because the ash is still 1.3 million years old! So for the "footprints" to be 40,000 years old, the ash must have been redeposited from some earlier source, and been plastic at just the time of a paleomagnetic excursion.
I love Renne's reaction:
Professor Renne: "How did I know they were going to say that? There is a finite possibility that that is correct, but the probability is extremely low."
That's why I try not to disagree with geochronologists!
It would be one thing if there were clearly consistent footprint trails. But there aren't. Since it's the floor of a quarry, I expect eventually to hear someone come forward who knows how the marks were made.
Chaw joins poop in archaeology arsenal
Well, archaeology is set to receive a once-in-a-generation influx of interest from teenagers drawn to the allure of the past. I mean, from the new Indiana Jones movie, of course.
So what do they have to go and do? Discover a real life crystal skull? Sorry, kids. If you want to be an archaeologist, it's all bodily functions from here on out.
Tom Dillehay and colleagues (2008) report in this week's Science that they have found chewed-on seaweed "cud" from Monte Verde, dated to 14,000 calendar years BP. And that paper is right next to the final publication of the Paisley Caves coprolites from Oregon, also dating to slightly before 14,000 calendar years BP.
Both papers are pretty cool -- together they emphasize that these kinds of forensic evidence are becoming increasingly important in documenting the activities (and existence) of archaeological populations. After all, a person has to poop thousands of times during his life, but he has only one skeleton.
Dillehay and colleagues interpret their seaweeds as a specialized medicinal collection, based on the presence of non-edible species and species present at different times of the year. Here's a quote from Michael Balter's news piece on the find:
Back 14,000 years ago, Monte Verde was located about 90 kilometers east of the sandy Pacific coast and 15 kilometers north of a rocky-shored inland marine bay. Algae from both environments were recovered, including inedible species that are today used as medicines in Chile and elsewhere. Moreover, the algal species found are known to flourish at different times of the year, suggesting to Dillehay's team that the Monte Verdeans were intimately familiar with coastal resources--possibly because they had originally arrived in the region via that route. Erlandson agrees: "The variety of seaweeds implies a pretty deep knowledge of coastal ecosystems and a long history of exploiting them."
Well, that's pretty impressive, even if the seaweed were chewed up. And hey, my kids are much more interested in bodily functions than they are in crystal skulls. So maybe this will bring in new archaeologists after all!
References:
Balter M. 2008. Ancient algae suggest sea route for first Americans. Science 320:729. doi:10.1126/science.320.5877.729
Dillehay TD, Ramírez C, Pino M, Collins MB, Rossen J, Pino-Navarro JD. 2008. Monte Verde: Seaweed, food, medicine, and the peopling of South America. Science 320:784-786. doi:10.1126/science.1156533
Gilbert MTP and 12 others. DNA from pre-Clovis human coprolites in Oregon, North America. Science 320:786-789. doi:10.1126/science.1154116
New news in New World settlement
Afarensis has a post on Brazilian evidence relating to the origins of Native Americans (via Gene Expression). It's a good summary of recent work by Neves and colleagues, including the background to this:
In essence, Neves et al, is saying that Paleoindians were part of the expansion of H. sapiens out of Africa, whereas Native Americans represent a later expansion of specialized populations. I find this suggestion intriguing. There are skeptics, as you can see from the above quote from The National Geographic article. Powell is arguing that because America was populated by small populations, genetic drift would kick in and you would see a lot of variation between populations and that the Lagoa Santo populations fit within the pattern af variation seen in Native American populations. If that is the case, I would expect one or two populations to display morphology similar to Australians and Melanesians. Instead, what we have is a series running from southern Chile to Florida - and now Kennewick in Washington (which was not included in either study)- all of which display morphology similar to Australians, Melanesians and Polynesians.
The idea of a "generalized" form for early modern humans has been around for awhile now, applied not only to Paleoindians, but also the Upper Cave crania from Zhoukoudian, and -- by some -- early modern Europeans.
Taking the "re" out of repatriation
Writer Rachel D'Oro of the Associated Press reports on the repatriation of human remains from On Your Knees Cave, Prince of Wales Island, Alaska:
ANCHORAGE, Alaska - Human remains estimated to be more than 10,000 years old will be returned to southeast Alaska Tlingit tribes 11 years after they were found in a cave in the Tongass National Forest.
It's the first time a federal agency has conveyed custody of such ancient remains to indigenous groups under the Native American Graves Protection and Repatriation Act, U.S. Forest Service officials said Friday.
...
Vertebrae, ribs, teeth, a mandible and a pelvic bone were among the remains discovered in 1996 during a Forest Service archaeological survey for a proposed timber sale on northern Prince of Wales Island. The area is the aboriginal homeland for Tlingit tribes.
Prince of Wales Island is in the Alaska Panhandle, near the British Columbia border.
Some archaeologists, noting the island's placement on the coastal fringe of Beringia, have suggested that it may document some descendants of the earliest settlers of lower North and South America. In this model, people were able to bypass the Cordilleran ice sheet by skirting southward along the coast of the Pacific Northwest. The first Americans lived significantly earlier than this site, which dates at the earliest to around 10,300 radiocarbon years. But the location makes the remains possibly the best skeletal representation of the original inhabitants of the Americas.
Rex Dalton described the site briefly in a Nature news article from 2003:
In the north of Prince of Wales Island, he ventured into the 'On Your Knees' Cave -- so named because it can only be entered by crawling through a narrow tunnel -- which turned out to contain a host of specimens. Subsequent years of summer digs by Heaton and his archaeologist colleagues have unearthed the oldest signs of human habitation in the Pacific Coast region, from sediments in the cave's floor.
The specimens include a bone tool that has been radiocarbon dated to 10,300 years ago, a human bone dated at 9,200 years old, and blades made from obsidian -- a volcanic glass found in lava beds -- of the same vintage. The latter have shown the ancient inhabitants of Prince William Island to be relatively well-travelled coastal seafarers. Craig Lee, an anthropology doctoral student now at INSTAAR, has used X-ray fluorescence spectroscopy to examine trace elements in the obsidian, and so determine its source. His studies show that the material in On Your Knees Cave came from Suemez Island, about 150 km to the south, and Mount Edziza, nearly 400 km to the northwest in mainland British Columbia.
Long-distance contacts along the coast, and between coastal and inland locations, among the early descendants of the New World invaders show the potential for rapid movement down the western coast of the Americas.
That archaeology expresses how important the site is for understanding early New World populations. But what about the skeleton -- how have the human remains contributed to our knowledge?
These remains are relatively fragmentary and therefore don't provide a lot of anatomical information that could be compared to recent people. But they contain a wealth of genetic information, of the kind that has been retrieved more and more routinely during the last 10 years.
Kemp and colleagues (2007) examined the DNA from the human remains, successfully extracting both mtDNA and Y chromosome markers that permitted analysis of the specimen's place relative to New World diversity. The conclusions help to emphasize the importance of such early skeletal remains:
Mitochondrial and Y-chromosome DNA were analyzed from 10,300-year-old human remains excavated from On Your Knees Cave on Prince of Wales Island, Alaska (Site 49-PET-408). This individual's mitochondrial DNA (mtDNA) represents the founder haplotype of an additional subhaplogroup of haplogroup D that was brought to the Americas, demonstrating that widely held assumptions about the genetic composition of the earliest Americans are incorrect. The amount of diversity that has accumulated in the subhaplogroup over the past 10,300 years suggests that previous calibrations of the mtDNA clock may have underestimated the rate of molecular evolution. If substantiated, the dates of events based on these previous estimates are too old, which may explain the discordance between inferences based on genetic and archaeological evidence regarding the timing of the settlement of the Americas. In addition, this individual's Y-chromosome belongs to haplogroup Q-M3*, placing a minimum date of 10,300 years ago for the emergence of this haplogroup.
Possibly the most significant aspect of ancient specimens is their ability to inform directly about the rate of sequence changes over time. Phylogenetic rates of change in mtDNA (e.g., between humans and chimpanzees) appear to have been relatively slow compared to the observed number of mutations between parents and offspring. There are two reasons for this: Some mutations are lost because of purifying selection, and some recur again and again at the same site, and therefore saturate over long time scales. But it is difficult to estimate the relative importance of these factors, and therefore it is hard to assign rates across different time scales (e.g., the founding of New World populations, modern human origins, etc.). Kemp et al. (2007) used the ancient mtDNA sequence from On Your Knees Cave to arrive at an estimate of the rate of change across the 10,000-year span, which ought to allow a more precise chronology for the New World founding:
Applying our most conservative rate of 34%/site per myr (95% CI 15-53%/site per myr) to the nucleotide diversity estimate (pi = 0.86) for mtDNA haplogroups A, B, C, and D in Native Americas (Bonatto and Salzano, 1997b), indicates that human entered the Americas ca. 13,438 YBP (95% CI 8,113-28,667
YBP). While this estimate does not preclude the possibility of an early entry, the estimate is also compatible with an entry more recent than 15,000 YBP (Kemp et al. 2007:617).
That is in contrast to some earlier genetic estimates, which had suggested founding times as early as 40,000 years ago -- often suggested as compatible with a very long-term occupation of Beringia.
Soon, we will surely see more testing of early specimens to try to assess the genotypes for other loci. For example, the idea of an early population bottleneck during the founding of New World populations has become an important hypothesis for explaining the limited variation of many genes in the Americas. But later events may also have been important in limiting variation. For example, Native American populations have substantially less allelic variation at HLA loci than do Asian populations. This reduction in diversity probably began with the initial founding and subsequent dispersal, but may also have involved natural selection during the last 10,000 years. The importance of the initial bottleneck may be tested by finding HLA types for early American skeletons -- if the founding event was almost solely responsible for the current HLA diversity, then the early skeletal samples should be comparable in diversity to recent samples. If they show much more diversity (for example, HLA haplotypes absent in recent Americans) then we can conclude that selection in later populations played more of a role -- implying a different scenario for pathogen-host evolution over time in the Americas.
In that light, repatriating a 10,000-year-old skeleton must subtract from our future inquiries into the origins of New World peoples. Or I should say, patriating, since it isn't being sent back to its people, it is being given to entirely new people who have no demonstrated relationship to the skeleton at all.
I do perceive the real benefits from cordial and cooperative relationships with indigenous peoples, particularly where tribes have well-defined and recognized historic territories. The vast majority of skeletal remains of ancient Americans are quite recent, and might in most cases (given sufficient evidence and analysis) be attributable to particular historic cultural groups.
But for remains over a few thousand years old -- and certainly for the earliest New World populations -- every living person of Native American descent may count these early skeletons among their ancestors.
References:
Dalton R. 2003. Archaeology: The coast road. Nature 422:10-12. doi:10.1038/422010a
Kemp BM and 13 others. 2007. Genetic analysis of early holocene skeletal remains from Alaska and its implications for the settlement of the Americas. Am J Phys Anthropol 132:605-621. doi:10.1002/ajpa.20543
Pig extinctions in Polynesia
I happened across an interesting article from last year by Christina Giovas that looks at pigs in Polynesia. People carried pigs with them to most of the islands that they colonized, excepting some distant and relatively recent places such as Rapa Nui (Easter Island) and New Zealand. These places presumably didn't get pigs, either because sufficient numbers did not survive the initial ocean voyage, because of low rates of subsequent contacts that might have brought pigs, or because the necessary forage for pigs was not present in sufficient quantity to establish them -- especially since pigs and people compete for many of the same foods.
But some islands that didn't have pigs at the time of European contact nevertheless have ancient archaeological evidence of pigs. So people brought pigs to those islands and established them, but they subsequently died out.
Here are the key paragraphs form Giovas' introduction:
Nonetheless, archaeological data in combination with the accounts of early European explorers reveal a pattern of pig distribution far more extensive in prehistory than at the time of historic contact (Fig. 1) (Alien et al. 2001; BayPetersen 1983; Bellwood 1987; Dye and Steadman 1990; Kirch 1991, 200Ob; Kirch and Yen 1982; Rolett 1998), suggesting that prehistoric Polynesians either allowed swine herds to die out or intentionally exterminated them. Whichever of these is the case, it appears that pig extirpation in Polynesia constitutes one of the few known instances of domestic species extinction (Ramis and Bover 2001), making it highly significant to the history of human-domesticate interactions.
It has been suggested that animal husbandry was simply less feasible on resource-impoverished islands (Anderson 2001, 2002; Bay-Petersen 1983; Kirch 2000b), particularly since husbandry practices involved feeding pigs cultivated crops, setting up an element of resource competition between pigs and their Polynesian keepers. Underlying this explanation are assumptions about island ecology and its impact on pigs and humans alike. Although the relationship is seldom made explicit, these ecological factors are presumed to operate according to principles of island biogeography, particularly the relationship of island area to species extinction. Other factors may also influence the likelihood of extinction, including island geology, elevation, and latitude and longitude. Here I test the relationship between pig extirpation and these variables. I propose that the pattern of prehistoric distribution for domestic pigs in Polynesia may be best understood in island biogeographic terms and that resource competition may represent the overarching causal mechanism driving pig extirpation, as Kirch (2000b) has argued. To begin, I offer a brief overview of the role of pigs in Polynesian society, followed by a discussion of the relevant principles of island biogeographic theory and the mechanisms-both environmental and cultural-that may have brought about pig extirpation in Polynesia (Giovas 2006:69-70).
Obviously, on the islands where pigs were extirpated, humans were a factor in killing them. But there is this question: did extirpation sometimes result from a deliberate plan, where people recognized that the pigs were eating vegetable foods the people needed themselves? Or did they just eat all the pigs, possibly during a shortfall, and not restock them from other islands.
In times of resource shortfall, pigs would have been effectively competing with people for the same agricultural produce, placing Polynesians in a position in which the benefits of pig husbandry may have been outweighed by its relatively high economic (energetic) cost. As Kirch and others have suggested (Alien et al. 2001; Bay-Petersen 1983; Kirch 2000b), trophic competition on this level may explain why animal husbandry failed in many parts of Polynesia. Those islands prone to resource shortfalls would have been at a greater risk for failure of pig husbandry simply because crop harvests may have been insufficient to support both humans and pigs. The onset of this effect need not have been sudden or severe. Instead, it may have operated slowly but systematically over decades or more (Giovas 2006:72).
Since many of the islands were in periodic contact with other societies, pigs might have redispersed in a number of cases, and possibly pigs were maintained over long times on some islands that otherwise would have lost them by recurrent contact. Giovas approached this issue of "social isolation" in an interesting way:
On the other hand, where human interaction between islands was frequent, it may have supplied a type of rescue effect (Brown and Kodric-Brown 1977) in which the continual immigration of domesticates safeguarded against their disappearance from an island. In this scenario, isolation comprises not only a geographic but also a social component in which substantial geographic isolation may be overcome in part by cultural forces. Weisler (1994, 1995, 1996, 1997; Woodhead and Weisler 1997), for example, has documented the existence of a long-distance economic interaction sphere among the remote islands of Mangareva and the Pitcairn group. The subsequent decline of settlements on Pitcairn and Henderson may have come about in part because of the breakdown of interisland voyaging between these islands (Giovas 2006:75, emphasis in original).
This concept of social isolation serving as a substitute for geographic isolation is very interesting to me; it adds a historically fluctuating component to the usual isolation factor treated by biogeographers.
Giovas found that island area was a strong predictor of whether pigs were extirpated, with most of the cases of pig loss being relatively small islands. That is in line with the prediction of island biogeography, at least in terms of the ability of larger islands to support more species within a given ecological space.
Island elevation was also related; extirpation was more likely on very level atolls. But it wasn't clear that this correlation remained true after controlling for island area; the two have a strong relationship. Still, a higher island might be one where humans have more trouble using all resources, as some remain at altitudes or on slopes used only occasionally by people. In other words, a higher island has more potential "wilderness" space where feral pigs might escape and avoid capture.
I find myself wondering whether an additional historical component may be important. Pig populations have the potential of growing quite a bit faster htan human populations, especially when human consumption of them is relatively low and natural forage is available. In other words, during the early years after a colonization, pigs may have grown faster than people.
But later in time, the people caught up in many islands; putting more and more pressure on the ecological space available to both people and pigs. In these contexts, the pigs would have lost out. So demographic growth and competition between humans and pigs predicts a pattern in which many islands start with pigs, but then lose them as human populations catch up and overtake them in size.
We might predict on this demographic basis that pigs would have died out on other, larger islands if Polynesian history had continued without interruption from European explorers. No island was likely to have been a steady-state with equilibrium in pig and human population sizes, although the joining of nearby islands into interacting culture groups might have moderated the effect.
References:
Giovas CM. 2006. No pig atoll: island biogeography and the extirpation of a Polynesian domesticate. Asian Perspectives 45:69-95.
Nobody but us chickens
Chickens were brought to South America in Precolumbian times by Polynesians.
It's all over the wires, so I'll just point to John Noble Wilford:
That is the conclusion of an international research team, which reported yesterday that it had found "the first unequivocal evidence for a pre-European introduction of chickens to South America," or presumably anywhere in the New World.
The researchers said that bones buried on the South American coast were from chickens that lived between 1304 and 1424. Pottery at the site was from a similar or earlier time. A DNA analysis linked the bones, which were excavated at El Arenal on the Arauco Peninsula in south central Chile, to chickens from Polynesian islands.
About half the news articles seem to be leading with, "Why did the chicken cross the Pacific Ocean?" This one is no exception. Har, har, groan.
I don't really have much to say about it, it seems like a very straightforward discovery. There is a radiocarbon date and the DNA evidence. Also, the early postcontact historical record corroborates:
Others thought [chicken arrival with the Spanish] unlikely, noting that when the Spanish invaded Peru in 1532, they saw chickens being used in traditional ceremonies. It seemed hard to believe, some scholars pointed out, that chickens would have been so rapidly dispersed from the east coast to the west and already be incorporated in religious events.
Charles Mann's book 1491 doesn't review this example, but it does talk about analogous kinds of early post-contact documentary evidence from New World peoples. It seems to be an increasingly important source of information to corroborate archaeological findings.
And the chicken transfer does help to explain the sweet potato problem:
The presence of the South American sweet potato in pre-European sites in Polynesia indicated some prehistoric contact between America and the islands.
I've never understood why this has been so quickly waved away. I mean, just last month this Nature News article reviewed a study that claimed that sweet potato seeds could have traveled to Polynesia on a free-floating unmanned ship.
Good grief. I suppose the chickens manned the boat. Er, chickened the boat?
Foods like these have to be among the fastest things to diffuse, even given relatively rare population contacts -- and certainly the Postcolumbian exchange between New and Old Worlds demonstrates this. I suppose the most negative evidence is the lack of evidence for other South American crops, like maize. Maybe the sea travelers used it all to feed the rats?
I mean, rats made it as far as Easter Island, why not the next step? Maybe some South American rats are actually Precolumbian also?
The only thing faster is disease -- and fortunately for the Americas, eastern Polynesia didn't have many of those to transfer.
Wilford's article ends with this:
Scholars found it disappointing and puzzling that the Polynesians who landed at El Arenal left nothing more than chicken bones. Pottery at the site is in a local style. Perhaps the visitors ate and ran, but not without leaving behind some starter chickens for future plates of arroz con pollo.
This seems quite obvious: the chickens were introduced a lot earlier than 1304. They would perhaps have been the longest-lasting element of the cultural transfer -- they wouldn't even have necessarily required permanent settlements, although a more substantial presence might well have occurred. Chickens are the most sensitive sign we can expect to see, just as sweet potatoes are the most sensitive sign on the other side of the contact. You aren't going to find the one place where Polynesians showed up -- you are going to find some site 500 years later when the chickens have spread across half the continent.
On Greenland Norse and Neandertals
I just read a good popular article by Colin Woodard about the 15th-century decline of the Greenland Norse.
"During the same time period, a lot of Norse settlements in Iceland and northern Norway were being abandoned, but nobody writes big books about that," [forensic anthropologist Niels] Lynnerup says. "I'm not sure that the Norse saw Greenland as being very different from the fjords they came from in Norway, and leaving it was no more stressful than abandoning a hamlet in Norway." His theory: In the 1300s and 1400s, Greenland's youths voted with their feet, leaving until the colony could no longer support itself. The last few left.
You may have seen some of this stuff before -- I think there has been a television special about it -- but it's a nice summary of what forensic anthropologists and others have been up to in southern Greenland, complete with analyses of livestock parasites.
I'm pointing to the article because of this line:
Did the Norse colonists starve? Were they wiped out by the Inuit - or did they intermarry? No. Things got colder and they left.
This is what I always keep in mind when I'm reading about "climate change killing the Neandertals" and whatnot. Human movements are fast. Hunter-gatherer migration is potentially much faster than migration of sedentary agriculturalists. In the archaeological record, we are not looking at migrations, we are looking at long-term fluctuations in the pattern of short-term, rapid movements.
There is no great dramatic moment. For the Norse, it looks like there were several bad winters. As the story relates, some people died, but most left.
The possible difference in the Neandertal case is the vastly longer timescale, which may have allowed a much greater cultural shift. It looks like the Norse did not mate with the indigenous Inuit in a way that left a lot of genes behind in Greenland natives. Their culture never adopted elements of Inuit subsistence technology that might have worked for them. Neandertals had longer to adjust -- and there were possibly fewer technological differences between them and their contemporaries in the first place. Cultural interactions seem to have occurred, and we can imagine mating would have also.
But in the end, climate change in Europe 35,000 years ago really amounted to a long run of bad winters. The people should have moved. They did move. Whoever was making Mousterian tools eventually stopped. As to what else happened, we have to rely on comparisons of other, later samples.
UPDATE (2007/01/06): A reader writes:
There is, of course, another comparison, the ESKIMO populations, who did not become extinct, or were even affected. The Norse were too arrogant to learn from the local "savages", not even Christians.
Good point.
Also, Razib picks up the theme.
Acheulean endings
There is no hard endpoint to the Acheulean; its tool types -- in particular the handaxe -- last well into the MSA/Middle Paleolithic. Here are some notes on the later occurrences in Africa:
Sheppard and Kleindienst (1996) considered the technological changes from ESA to MSA at Kalambo Falls, finding:
...there is little change, at this site, in the basic techniques of blank production or the attributes of the blanks produced from the ESA to the MSA. The only marked change to occur is the loss of large cutting tools (hand axes, cleavers) and their replacement by heavy-duty forms (core axes, picks). It is hypothesized that this change marks a decline in portability as a factor in the design of large edge tools (Sheppard and Kleindienst 1996:171).
McBrearty (2003) reflects on the late persistence of characteristic Acheulean handaxes in association with some of the earliest fossils of "modern" humans, the Herto hominids from the Middle Awash of Ethiopia:
At 160-154 ka, the Herto handaxes are the latest survivors of Acheulean technology known in Africa. Formerly, the best candidates were those from the site of Rooidam, South Africa, where a U-series date indicated they might date to ~ 170 ka (Szabo and Butzer 1979). K/Ar dates of ~ 240 ka for Acheulean artifacts from the Kapthurin Formation, Kenya (Leakey et al. 1969, Tallon 1978) have frequently been cited as the date for terminal Acheulean in east Africa, but new more precise 40Ar/39Ar dating now shows these to predate 285 ka (Deino and McBrearty 2002).
What does the presence of Acheulean tools at 160 ka signify? The nature of archaeological change in the African Middle Pleistocene is murky, but the most obvious development is the abandonment of Acheulean technology, usually thought to have been made by H. erectus and its replacement by implements of MSA traditions, believed to have been produced by H. sapiens. In a nutshell, this comprises the replacement of handaxes and cleavers by points, signifying a shift from hand-held to hafted tools, and the birth of projectile technology. The presence of Levallois technology is usually considered to be a feature of the MSA, but it is present in late Acheulean contexts as well. In fact, the big flakes used to make African late Acheulean handaxes, including some Herto and Kapthurin specimens, may be struck from Levallois cores....The earliest MSA points were once thought to be those from Gademotta, Ethiopia, dated by K/Ar to 235 ka (Wendorf et al. 1994). Retouched points from the Kapthurin Formation predate those from Gademotta by 50,000 years. Dated by 40/Ar/39Ar to > 285 ka (Deino and McBrearty 2002), the are currently the oldest in Africa. A similar date is estimated for the basal MSA at Florisbad, South Africa (Grün et al. 1996), though these levels have not yet yielded points (Kuman and Clarke 1986, Kuman et al. 1999). Both retouched points and Levallois points have been recovered at Twin Rivers, Zambia, where they are dated by U-series to 265 ka (Barham 2000) (McBrearty 2003:1-2).
McBrearty briefly discusses two alterative hypotheses. One is that the Acheulean handaxes are markers of a biological species, and that MSA points and other artifacts were made by some other species, or multiple species. The problems with this hypothesis are manifold, including the association of the handaxes with fossil remains that would be uncomfortable fits in H. erectus at best. Yet, the hypothesis can be sustained by the additional assumption that multiple species cooccupied the same African paleolandscapes with sufficient contemporaneity to drop their tools in the same places:
Or can the archaeological record of the African Middle Pleistocene represent the behaviour of several species? Might different tools be the distinctive signatures of different populations or species who are competing with each other for the same territory and resources? A similar scenario of interspecific competition seems to have been played out between H. sapiens and the Neanderthals in Europe after 40 ka, but there the replacement was complete within 10,000 years or so. Africa is a far more enormous continent than Europe, its Pleistocene population far less dense, and our knowledge of its record less finely resolved. The 125,000 years represented by the Acheulean to Middle Stone Age transition in Africa is more than four times greater than the entire timespan of the residence of H. sapiens in Europe, but only a handful of African sites from this long time period have yet been explored.
The other hypothesis is that the handaxe appears wherever it is functionally relevant, regardless of phylogeny. Acheulean elements may have been independently invented again and again:
Archaeologists routinely stress the uniformity of the Acheulean tradition, and rarely entertain the possibility that the category 'handaxe' is an artificial construct that we impose upon the material. Experimental work by Jones (1994) shows that the familiar tear-drop shape of the handaxe is a design compromise resulting from attempts to gain the maximum cutting edge from the minimum weight of stone. The same techniques of blank production and bifacial retouch were employed in many cases to make both Acheulean handaxes and MSA points, and in the Kapthurin Formation the two classes of artefacts can grade into each other. Furthermore the tear drop shape occurs repeatedly in the archaeological record in cases where there is no evidence of 'phylogenetic' relation. Many projectile points from Africa, Europe, the Near East, and even the New World have similar plan forms (McBrearty 2001; Otte 2003), but there is clearly no 'phylogenetic' relationship among them. Rather, the recurrence of this target form demonstrates the design constraints inherent in fractured stone, and it appears to have been rediscovered repeatedly by tool makers in the course of prehistory (McBrearty 2003:2-3).
Of course, if the recurrence of the handaxe across the 1.4 million years of the Acheulean resulted from its function rather than a continuous chain cultural transmission, then it takes away one of the principal pieces of evidence that Acheulean people had any substantive culture at all.
As far as handaxes are concerned, that may be a good thing. The maintenance of a single cultural tradition across much of three continents over a million years by exclusively social transmission seems incredible. Some have suggested that the handaxe is hardwired into the human genome, a proposition that seems even less credible (at least, to me). Absent these means of transmission, we are left with the proposition that the handaxe did not fade from the earth because of its functional utility -- either it was the tool that did the job the best, or it was the best tool that humans were capable of making that did the job adequately. If function and human ability combined to make it so persistent, it should not be the least bit surprising that the form should recur in later contexts -- in the scenario of recurrent invention, it never really went away. The production of handaxes on large Levallois flakes seems especially relevant.
One view of the uniformity of handaxe variation across space is that humans maintained strong cultural links across large distances. The knowledge of handaxe production and the concept of handaxe shape were highly conserved cultural markers that were widely spread.
But the alternative hypothesis, that handaxe form was repeatedly invented by hominids and actually served as the best possible tool given its purpose and hominid flaking capacities, does not require long-distance cultural interactions to maintain the uniformity of the tradition. Nor does it require the recurrent movement of people over vast distances to spread the information. Given the lack of evidence for such long-distance movement, this hypothesis gains a bit of credibility.
Raw material utilization in the Acheulean is highly localized. People may have been using the large tools as repositories for further flaking, and they may have curated them for some time. But they did not transport them over vast distances. Ordinarily, Acheulean people do not appear to have transported material over distances greater than the possible daily movement of a single individual -- a few tens of kilometers.
It raises an obvious question: Is high mobility (in other words, hunter-gatherer-like mobility) a shibboleth of the assumption that handaxe form required culture? Did earlier humans actually have very small, limited home ranges and daily movement patterns? After all, if humans carried handaxes around for their utility, and they didn't manufacture a new one every day, then some handaxes ought to represent week-long movements or even longer.
References:
Sheppard PJ, Kleindienst MR. 1996. Technological change in the Earlier and Middle Stone Age of Kalambo Falls (Zambia). Afr Archaeol Rev 13:171-196. DOI link
McBrearty S. 2003. Patterns of technological change at the origin of Homo sapiens. Before Farming 3:1-6.
How monolithic was the Acheulean?
I'm taking some notes on change and stasis during the Acheulean, and they're not entirely complete, but in the interest of clearing my desktop I'm going to start posting them in sections.
We can start with some epitaphs on the era as a whole:
Standard textbook knowledge informs us that the Acheulean was remarkably the same throughout its geographic distribution. For example, African handaxes are supposed to be very similar to European and Indian handaxes. It is also well known that handaxes come in a variety of shapes. However, there seemed to be no clear correlation between the various shapes and the geographic regions in which they were found, giving the Acheulean a markedly monotonous flavor (Wynn and Tierson 1990:73).
Currently, two perspectives dominate the interpretation of these tools. The first is that they result from "mental templates" held by the knappers and reflect shared cultural norms. The second is that their shapes were an unintended result of a knapping procedure and the use life of the artifact and have no implications for culture or cognition (Wynn 2004:674).
Regional differences
One argument favors cultural continuity as an explanation for handaxe persistence: there appear to be some regional differentiation of modal handaxe form in the later Acheulean. Wynn and Tierson (1990) find that handaxes from the later Acheulean of Israel had an average form that was wider than those from Africa, England, or India. Comparable differences in form did not distinguish these other regions from each other. Wynn and Tierson (1990) concluded that cultural transmission was the most reasonable explanation for this interregional difference. The difference in form was consistent across different raw materials, and they considered raw material a confounding variable contributing to variation, rather than an explanation for the interregional difference. They could not eliminate the possibility that the difference reflected temporal change -- since most of the sites were effectively undated.
Regional differences in form are not a problem to explain under the noncontinuous account. As handaxes were independently invented in several different traditions, there is no reason to think they should be exactly alike. Within the functional constraints upon them, there was evidently substantial room for variation -- as indicated by the variation of form within regional samples, and within individual sites. Undoubtedly, there was substantial variation in form among handaxes crafted by individual makers.
What may be surprising is that there was not more interregional variation in form. This uniformity may appear as evidence for cultural or genetic transmission. But it may equally represent the lack of transmission. If different local groups or regional populations did not share a long continuous genetic or cultural chain of handaxe-making instructions, then they would have effectively explored the functional space of bifaces repeatedly in each region. This exploration would have introduced substantial temporal differences between levels at single sites. Different regional samples would each represent samples of the same, broad, statistical population of handaxe variants.
And of course, if the function of handaxes was really to serve as a core for the production of flakes (discussed here), then there is another good reason why regions might vary in handaxe width in particular -- some local regions may have had more intensive raw material utilization than others, regardless of the raw material type. Where exposures were common, people would have been less likely to use the same cores as intensively. Where cores were rapidly buried, they would have been less likely to reuse them to full depletion. This is essentially the explanation favored by McPherron (2000) for the same pattern of observations: the differences may be explained by a combination of raw material size and reduction intensity. In support of this idea, he notes that Gowlett and Crompton (1994) show that handaxe width is allometric against length; smaller and more intensively flaked handaxes tend to be relatively shorter -- and hence, relatively wider compared to their lengths.
Standardization and mental categories
McNabb et al. (2004) examined assemblages of bifaces from ESA sites in southern Africa. They found that standardization and symmetry among the bifaces were "not a high priority". In contrast to the hypothesis that handaxe form is a mental template imposed on raw material, they concluded that there was little evidence for standardization, beyond the size and general functional characteristics of the bifaces.
From this lack of standardization, McNabb et al. (2004) speculate that the Acheulean people did not learn to make tools by internalizing a strict concept of form, and did not use tool shape itself to convey information.
Rather than being a tightly conceived and culturally sanctioned outline form acquired and maintained through social learning, the shape of a large cutting tool was a variable idea in the mind of the knapper. The specific form of the tool depended on individual ability, the perceived constraints imposed by raw-material shape, function, time, and social factors, all of which could change with time, place, and circumstances (McNabb et al. 2004:667).
McNabb et al. (2004) proceed to argue for the idea that these tools were maintained as individualized memic constructs, meaning that the idea of making the tools was socially learned, but the final shape of the tools itself was not the object of that learning. Certain steps in tool production may have been conceptually standardized, meaning that the toolmaker learned certain tricks through observation and carried those out in a standardized fashion, but did not standardize the end product.
Standardization is not necessary to the physical function of a stone tool. The presence of standardization implies a non-physical function; either communicative or normative. The claim that Acheulean bifaces were extensively standardized to a common template is a claim for a nonphysical role -- standardization cannot be explained in the absence of such a function.
So a claim that Acheulean bifaces are not standardized is significant in that it denies a nonphysical (i.e., symbolic or communicative) role for these tools.
McPherron (2000) summarizes much of the 1990's debate about regional and temporal differences in biface manufactures, and draws out the conclusion of functional rather than mental uniformity in two parts:
It still must be questioned, however, whether these data actually indicate "mathematical transformations" or "mental templates" as Gowlett suggested. Is it really surprising, for instance, that length and width converge on zero? The fact is that in a reduction technology such as lithics, it can hardly do otherwise.... The only way it could be otherwise would be if the length or width removal rate changed at some point in the reduction sequence. If this were the case, then length and width would not be highly correlated, and Gowlett would not be able to argue for a high level of standardization (McPherron 2000:662, referring to Gowlett 1984).
Moreover, the very distinct possibility that handaxes were reworked and that shape changed as a result, challenges the basic underlying assumption of most studies of handaxe shape: namely, that a particular shape is the desired end-product of the bifacial reduction sequence. This problem exists regardless of whether handaxes were cores or tools or both. Either way, handaxe shapes represent stages in a continuous reduction process. Given the data presented and reviewed here, it can be concluded that this reduction process was remarkably similar in various parts of the Old World and at various times in the Middle and Late Pleistocene, suggesting that if there were a mental template, it was the same mental template (McPherron 2000:662-663).
In contrast, Wynn (2002) suggests that bifaces -- at least, those produced late in the Acheulean -- were products of people who possessed mental categories for them. Wynn (2002) outlines and describes evidence for attention to three aspects of symmetry in stone tools, beginning with congruency and proceeding to three dimensional symmetry and finally the deliberate breaking of symmetry.
These handaxes were almost certainly categories, and categories are abstract, multi-modal, and rely on associative memory. As such they reside in declarative memory, which "requires associative links between several types of information that are stored in different areas" (Ungerleider 1995, p. 773). These hominids could manipulate perspectives and spatial quantity, produce congruent symmetries, and even distort these principles to achieve striking visual effects (Wynn 2002:397).
I am ambivalent about whether and to what degree such artefacts may have had a social role beyond their utilitarian function. I think it unlikely that hominids who could deliberately alter symmetry to achieve an effect would not also apply this mastery upon some other artifact. If these modes of manufacture were learned socially, and if they could vary arbitrarily from a single template, then there should have been not only individual variations (which have a more-or-less constant distribution across space) but also regional and local variations (which would have a discontinuous distribution across space). Yet there is precious little evidence for such interregional variation -- or "style" -- until the MSA and Middle Paleolithic. Even there, it is not obvious that the regional differences are stylistic rather than functional. There are suggestions that other artifacts expressing "categories" did exist in the later Acheulean, including possibly carved sculptures like Tan Tan and Berekhat Ram, and deliberately incised lines like the Tata artifact. More of these would substantiate the case better. As it is, these are credible as products of individuals, but do not document a communicative aspect to artifact production or a tradition of such social and mental categories.
On the other hand, the null hypothesis of "no mental templates" must surely apply with more force to the earlier part of the Acheulean, and less to the later part. Wynn's (2002) argument may well apply with force to the later Acheulean, the period he examines, while not being true of earlier humans. In this case, the evidences for manipulating symmetries in artifacts may substantiate an increasing technical skill and "concepts" in some sense. The communication of a concept through the use of external objects would indeed represent an advance in human mental interactions and social learning. In this context, a reading of Wynn's (2002) response to discussants on his paper is worthwhile. In particular, he makes a clear distinction between earlier and later Acheulean in terms of the recognition and manipulation of symmetry.
References:
Gowlett JAJ, Crompton RH. 1994. Kariandusi: Acheulean morphology and the question of allometry. Afr Archaeol Rev 12:3-42.
Sheppard PJ, Kleindienst MR. 1996. Technological change in the Earlier and Middle Stone Age of Kalambo Falls (Zambia). Afr Archaeol Rev 13:171-196. DOI link
McNabb J, Binyon F, Hazelwood L. 2004. The large cutting tools from the South African Acheulean and the question of social traditions. Curr Anthropol 45:653-677.
McBrearty S. 2003. Patterns of technological change at the origin of Homo sapiens. Before Farming 3:1-6.
Wynn T, Tierson F. 1990. Regional comparison of the shapes of later Acheulean handaxes. Am Anthropol 92:73-84.
McPherron S. 2000. Handaxes as a measure of the mental capabilities of early hominids. J Archaeol Sci 27:655-663. DOI link
Wynn T. 2004. Comment on "The large cutting tools from the South African Acheulean and the question of social traditions". Curr Anthropol 45:672-673.
Wynn T. 2002. Archaeology and cognitive evolution. Behav Brain Sci 25:389-438.
What won't a handaxe do?
I've had this working paper by Tony Baker on my desktop for awhile, and it has been discussed on some message boards. I wanted to link before I forget. It's a good online review of the problems interpreting Acheulean handaxes. Baker's preferred theory is that the "handaxes" were principally cores for flake extraction. This is not a novel view, but the unique aspect is the way he derives the argument from the physics of flaking:
Homo erectus did not select small cores from which to extract flakes (make handaxes). I propose he chose large cores because he did not have the manual dexterity to externally support them and, therefore, he had to rely on the inertial support. He just let them lie flat in his hand or on his leg. This inertial support knapping meant all blows had to be directed toward the center of mass. Flakes scars could not pass the center of mass so the handaxe remained relatively thick. Additionally, with numerous flake removals the handaxe became smaller and more ovate or discoid in shape because the blows were being directed from the furthest edge from the center of mass toward the center of mass.
This is all described well, which is what makes the paper worthwhile. This paragraph is a good encapsulation of the problem of the traditional view:
The Lower Paleolithic researchers who believe the Acheulean handaxe was the desired product do not find their justification in its function, since its function is not understood (Bordes 1968:64; Debenath and Dibble 1994:130; Gowlett, Crompton and Yu 2001:612; Isaac 1977:12,144; McPherron 2000:73; Roe 1981:271). Instead, they find their justification in its unchanging morphology. For a million plus years, its basic shape remained constant as it spread across three continents. It has a tip, a butt, and often symmetry in three dimensions. Its length is rarely longer than twice its width, and Gowlett has demonstrated at Kilombe, that a linear relationship between length and width explains 69% (R=0.83) of the variation (1995:202). Issac proved there was a strong correlation between the length/width ratio and length regardless of sites or continents (1977:139). Further, its width is rarely greater than three times it thickness. This unchanging morphology is, therefore, proof that the handaxe was constructed to conform to an unchanging, deeply engrained mental template.
When I looked at this first, I thought I would post an expanded review of the handaxe problem, but other matters delay that at the moment. In the meantime, Baker's paper is a valuable resource.
In summary, my belief is that the handaxe was a large biface core made by Homo erectus individuals who lacked the manual dexterity of modern humans. I arrived at this conclusion based on the morphology of the handaxe. And, since there is no fossil hand evidence to suggest otherwise, my theory is intact. I am forced to admit that I believe Homo erectus' hand was evolving during its 1.6 million years of existence. Therefore, I would expect there also was an evolution in the handaxe from large to small and thick to thin. Yet, this evolution is only going to be visible in vertically stratified sites or well-dated sites. And, then it will still be difficult to identify because the handaxe was the by-product of expedient flake extraction. It was not the desired product.
I think it would be better to replace "hands" here with "brains", as the prerequisite "dexterity" in this context probably is more mental than manual.
Technological sophistication of the earliest toolmakers
One study reported at the meetings brought to mind a growing literature on the sophistication of Pliocene archaeological assemblages, grouped as "Oldowan" or "pre-Oldowan." Most Oldowan tools lack a standardized form, and the assemblages are dominated by flakes, core tools with one sharp edge or point, and unmodified manuports or hammerstones. The lack of standardized tools (such as handaxes) lead easily to the conclusion that Oldowan technical assemblages are simple, so that anyone could make one who could chip rock.
But a number of recent sources illustrate that this perception isn't accurate. The Oldowan encompasses a variation of manufacturing complexity and other elements, but cannot be said to be merely rocks that have been carried or bashed together. The most interesting studies have examined the issue of raw material exploitation and the actual processes used to reduce stone into useful tools.
David Braun gave a presentation that relates to this topic at the 2005 Paleoanthropology Society meetings. In an analysis of the lithic remains at the Kanjera South site (Kenya), the group found that the intensity of the reduction sequence varied with source material.
Kanjera South, dating to around 2.2 million years, preserves one of the largest early records of material culture. According to Plummer et al. (2001:810):
Hominins at Kanjera South utilized a wider variety of lithic raw materials than found at most Oldowan sites, some of which (chert, quartz, quartzite) must have been transported from outside the immediate vicinity of the deposits since they are not present in the local clast population.
In the study by Braun et al. (2005), fine-grained quartzite, which had a nearest source outcrop a long distance away from the site, was reduced more intensively than locally available rock. This reduction involved more flakes taken per core, and occasional retouch. They hypothesized that this difference was due to the hardness of the quartzite and its ability to hold an edge, which makes tools made on quartzite potentially more useful. The hominids apparently recognized these qualities in the raw materials and exploited the quartzite. We can probably infer that the transport distance was a result of this recognition as well.
Stout and colleagues (2005) report on the use of raw materials at Gona, which is at present the earliest site at which stone tools are found, dating to around 2.6 million years. This study sampled unmodified stone taken from conglomerates that represent natural accumulations. If stone were taken merely from local sources without selectivity, the modified stone assemblages from the site would be expected to match these conglomerates. As described by Stout et al. (2005:366-367), the past understanding of Oldowan toolmakers has suggested that they may not have exhibited a clear understanding of the technical qualities of their raw materials:
At Koobi Fora, the high frequency of basalt in local conglomerates (i.e. ancient stream channels) is closely mirrored by its predominance in the artifact assemblages (Toth, 1985; Schick, 1987; Isaac et al., 1997). Although quartz, chert and glassy volcanics are "reasonably easily available" in the conglomerates, these material types do not appear to have been specifically selected for (Isaac et al., 1997:268).
This paper gives an excellent review of the history of raw material choice. In brief, the literature review indicates that early toolmakers at Koobi Fora may have exercised some degree of selectivity by avoiding certain kinds of stone (vesicular lavas and weathered cobbles). By Olduvai Bed II (after 1.7 million years ago), some toolmakers were clearly going out of their way to acquire certain stone materials:
The increased use of chert at Olduvai around 1.65 Myr was clearly occasioned by the temporary exposure of rich sources of this material by the retreating waters of the paleo-lake (Hay 1976). Hominid toolmakers at this time readily appreciated the superior flaking properties of chert, leading to the formation of the earliest known special-purpose quarry site at MNK CFS (Stiles et al. 1974; Stiles 1998) (Stout et al. 2005:367).
So when did this technical knowledge really originate? Was it a skill that developed gradually over the million years between Gona and Olduvai Bed II?
Stout and colleagues find that the materials used for tools at Gona do not match the natural cobbles that were avaiable in a number of respects. In essence, the hominids ignored many kinds of stone that had fracture patterns or other properties that made them less suitable as raw materials. They (2005:368) state the conclusion most clearly at the end of their introduction:
The strong pattern of raw material selection seen at Gona demonstrates that low levels of selectivity are not universal in the Oldowan. The fact that the assemblages from Gona are the oldest known in the world also argues against an overarching temporal trend toward increasing selectivity within the Oldowan.
The flaking skill of early hominids is the subject of a paper by Roche and colleagues (1999) in Science. At the two localities of Lokalalei 1 and 2C, both dating to 2.34 million years ago on the west shore of Lake Turkana, this study accomplished refits of stone flakes and cores to directly inspect the reduction sequence. They found that this sequence was quite complicated in many of the instances of manufacture at the two localities:
At LA2C, the dominant reduction sequence consists of these more reduced cores. Unidirectional or multidirectional removals are flaked on a single debitage surface, from natural or prepared platforms. This knapping system allows the production of a long series of removals without changing the volumetric structure of the core. The repeated application by the knappers of the same technical principles to a whole series of cores, and during the reduction of each core, indicates an elaborate debitage scheme, implying motor precision and coordination. Thes