Read an Excerpt
By Kelly E. Graf, Caroline V. Ketron, Michael R. Waters
Texas A&M University PressCopyright © 2014 Kelly E. Graf, Caroline V. Ketron, and Michael R. Waters
All rights reserved.
Occupying New Lands: Global Migrations and Cultural Diversification with Particular Reference to Australia
In the Australasian region cultural differentiation, experimentation, and adaptation characterize the global dispersion of Homo sapiens. The migration of humans out of Africa and into Australia was not a singular process governed and guided by persistent traditions; normative and static images of social and economic organization cannot explain the diversity of cultural evidence associated with the dispersion. This paper reviews the evidence for a dynamic process of social, economic, and technological diversification associated with the spread of humans and their adaptation to new social and physical environments. The evidence is read as processes of adaptation, cultural connectedness, and isolation, interacting in complex and regionally varying ways. In the case of Australia the patterns of occupational history and cultural trajectories were not the product of conservative maintenance of social tradition or a diminished set of cultural practices following serial founder events. The evidence is best interpreted as a radiation of cultural practices and perspectives that reveals an adaptive response to the occupation of new lands.
KEYWORDS: Australia, Sahul, Colonization, Out of Africa, Adaptive radiation
Australia was colonized by groups of Homo sapiens whose ancestors had migrated out of Africa many millennia earlier.
Global dispersal of humans was a prolonged, complex process with multiple migrations in many regions. Archaeological and genetic evidence is currently revealing some of the complexity of that geographic expansion. Human settlement of Australia is a crucial case study of this dispersal process. The colonization of Australasia is one culminating event of the global expansion and must be understood in light of what we now know of the global process. At the same time Australia provides a landscape that illuminates the timing and nature of human dispersals, particularly in the region of Oceania and East Asia. Australia enhances our understanding of early migratory processes because of a rare combination of qualities: It was colonized by one of the first waves of migrating H. sapiens moving eastward from Africa, the land was previously empty of other hominins thereby giving clarity to the initial occupational period, and there is a growing quantity of good-quality archaeological data available for parts of the continent. In this chapter I summarize current evidence for the arrival and dispersion of humans across the Pleistocene Australian continent, called Sahul, in the context of what we know of the broader eastward dispersion within the Old World.
Current debates about the phase of colonization in Australia focus on four critical questions of evidence. First, when did humans arrive and at what rate did they settle the diverse environments across the continent? Second, did human colonizers contribute substantially to environmental transformations in Australia? Third, what was the technological repertoire of foragers at and following colonization; and was that repertoire narrowing over time? Fourth and finally, what does public signaling, revealed archaeologically through art or ornamentation or ritual, tell us about the configuration of the colonizing societies.
Answers to these questions have played a central role in discussions of not only the arrival of people in the Australian landscape, but more broadly in models of the dispersion process between Africa and Sahul. I will discuss key strands of evidence for each of those four questions.
Timing of Colonization and Rate of Settlement across Australia
Archaeological evidence indicates humans probably arrived in Australia between 45,000 and 60,000 years ago, and most likely somewhere around 50,000–55,000 years ago. This "window" of colonization is well established but imprecise, so that while some researchers prefer an antiquity in the low 50,000s others argue that humans arrived slightly later. These differences of opinions between Australian archaeologists are strongly held, and hence any statement about timing of colonization is considered controversial in a debate that increasingly shadows the Clovis/pre-Clovis controversy of the Americas. My view, as outlined in Hiscock (2008), is that
1) given the complexities and imprecision of both radiometric age estimates and stratigraphic associations the only statement that would be generally accepted today is that the earliest evidence for humans in Australia dates to 50,000 ± 5,000 years ago,
2) that given the extremely small sample of sites known the age of the earliest occupation identified thus far must be considered a minimum age for the colonization, and
3) the difference of a few thousand years may make very little difference to the process of arrival: either end of this "window" is long after H. sapiens exited Africa, and though the size of water barriers would change over time those variations need not have posed significant problems for colonizers.
Much of the debate about the possibility of pre-50,000 occupation rests on the interpretation of sites in western Arnhem Land, near to the modern northern coastline. There, excavations at Nauwalabila and Malakunanja II yielded layers containing artifacts in association with sediments that have been estimated to be between 50–60 kya. While some archaeologists have accepted these age estimates as indications of human antiquity in the continent, others have rejected the stratigraphic association between dated materials and cultural materials and have argued that colonization was more recent. The most strident critique of association, and the strongest advocacy of a more recent age of colonization, has been advanced by Jim O'Connell and Jim Allen in a series of articles over the last decade (Allen and O'Connell 2003, 2008; O'Connell and Allen 2004, 2007, 2012; O'Connell et al. 2010). They argue humans can only be reliably dated to 44,000–46,000 years BP (O'Connell and Allen 2012:10), and that claims for earlier occupation have been refuted.
Their claim is based on an interpretation that vertical movement within the Arnhem Land shelters, principally as a result of termite activity, has brought lithic artifacts and dated sand grains into a false association. In arriving at that conclusion they argue that all the lower artifacts have moved substantially downward through the sandy deposits and they imply that stratigraphic evidence that would document minimal vertical displacement, of a small pit dug more than 40,000 years ago claimed for Malakunanja II, is inadequately documented and perhaps that the pit may not exist (O'Connell and Allen 2004). These propositions are plausible, and O'Connell and Allen cite instances where substantial vertical displacement has been documented in Australian sandstone shelter deposits (e.g., Richardson 1992). And yet the evidence from Australian shelters also shows that where vertical movement occurs it is bidirectional, with artifacts moving up as well as down, and that substantial movement is not universal—some sandy deposits retain reasonably precise indications of the associations created at deposition (see Richardson 2011; Stockton 1973). O'Connell and Allen have not re-analyzed either the artifacts or the sediments recovered from Malakunanja II, and so while their models of taphonomic disturbance are relevant to the colonization debate they are not proof that artifacts had moved and are consequently not associated with the sediments well over 50,000 years old from which they were recovered.
O'Connell and Allen seek to bolster their case by arguing that the inability of archaeologists to find more sites older than 50 kya indicates that there are no such sites to be found (e.g., Allen and O'Connell 2003:17). However, in the absence of discussions about the kinds of sites that have been dug in recent years and the probability of finding sites of high antiquity (see Langley et al. 2011; Surovell and Grund 2012), this argument is not compelling. We might equally argue that finding one or two sites of that vintage was remarkable luck given the small sample of early sites we had and still have. In any case, Malakunanja has recently been re-excavated by Chris Clarkson and his analysis of the site will shed light on its formation and antiquity in the near future.
Whatever consensus emerges for the Arnhem Land sites there are a number of sites from the central and southern portions of the continent that are widely accepted as indicating occupation between 45,000 and 50,000 years ago. For instance, in southeastern Australia debris from human occupation was present at Lake Mungo, with the stratigraphically lowest artifacts bracketed by luminescence dates of 45.7 ± 2.3 and 50.1 ± 2.4 kya (Bowler et al. 2003). Although the stratigraphic integrity of these deposits have been queried (O'Connell and Allen 2004), the position of artifacts within the deposit has been coherently explained as reflecting ancient landsurface slope by the excavator (Shawcross 1998; Bowler et al. 2003), and there is little reason to doubt their association, leading researchers such as Smith (2013) to infer use of the palaeolake soon after 50,000 years ago.
In southwestern Australia the Devil's Lair cave was occupied before 46,000–47,000 years ago (Turney et al. 2001). A thick fan of alluvium in Layer 30 Lower is dated by radiocarbon at 44,000–47,000 BP. Artifacts and faunal remains are found far below this level, as far down as Layer 38. These layers are dominated by limestone rubble, which would have limited vertical movement of large objects. O'Connell and Allen (2004:840) emphasize that these layers have small channels indicating water flow, and they conclude that the artifacts contained within them may not be in primary deposition context—a point noted long ago for the bones (Balme 1980). Based on this concern O'Connell and Allen (2004:841) argue that
a human presence at Devil's Lair somewhere in the 41–46 ka range is now indicated, the argument for an earlier occupation remains equivocal.
That conclusion is unsustainable. I have examined the specimens from layers 31–38 and they are clearly artifactual, including a retouched flake, and the specimens are made on calcrete, which is uncommon in layers 30 and above. My observations agree with Dortch's characterization of the assemblage (reported in O'Connell and Allen 2004) and reinforce Dortch's point that the early assemblage in layers 31–38 is not likely to be derived from higher layers by vertical movement. Obviously the artifacts from the lower layers may have initially been deposited in sediments elsewhere within or near the cave, then eroded and redeposited into their levels, but if redeposition occurred it did so long before 44,000–47,000 BP and the artifacts in question are substantially older than that date. The conclusion of Turney et al. (2001), that Devil's Lair represents human occupation at close to or before 50,000 years BP, stands on current evidence.
A number of other sites currently appear to have material dated to about or more than 45,000 years BP, including Carpenters Gap in the northwest (Fifield et al. 2001), Nawarla Gabarnmang in Arnhem Land (David et al. 2011), and Parnkupirti in central Australia (Veth et al. 2009). Taken together with the information described above from Devil's Lair, Lake Mungo, Nauwalabila, and Malakunanja II, this suite of sites shows that human occupation of the continent, including southern regions, took place before the 45,000–46,000 BP date that O'Connell and Allen have hitherto accepted. However, given the imprecision of age estimates in many of these sites it is not currently clear whether the earliest evidence of occupation in these regions is slightly before or around 50,000 years BP or closer to 55,000 BP, and consequently the published interpretations (many of which ignore the statistical implications of error values) do not provide a basis for evaluating whether there was a difference in the antiquity of archaeological materials between different regions in the north or south of the continent. This has rendered models of the rate of dispersion across the Australian mainland largely untestable (see Rindos and Webb 1992).
An additional consideration is that these archaeological signals are unlikely to mark the arrival of humans on the continent and it is more probable that colonization occurred well before the dated occupation levels. One reason for this is that even with constant rates of site destruction there will be progressively fewer early sites preserved over time. This age-related process of site destruction is non-linear, as demonstrated by generic modeling (Hiscock 2008), archaeological datasets from Australia (Williams 2013), and non-archaeological sediment records from North America (e.g., Surovell et al. 2009). The cumulative consequence of attritional processes, continued over a prolonged period, would be that very few sites from the period of colonization would have survived until today. Recent arguments by Surovell and Grund (2012), using calculations of predicted attrition rates, indicate if there was a phase of megafauna extinction in Australia dated to more than 40,000 years BP it is unlikely that any megafauna kill sites would be preserved or discovered. This conclusion could, and should, be directly converted into a statement of how improbable it is that sites from the even earlier phase of colonization would be discovered, reinforcing the conclusion that maximum site age so far established underestimates the age of colonization.
Differential destruction of sites in different locations and landscapes has also been considered to have obscured the record of colonization. For instance, the Pleistocene coastline was hundreds of kilometers from the modern one, and substantial tracts of land that colonists presumably occupied have been lost to sea-level rise, potentially destroying the occupational debris of many generations of settlers—an argument rehearsed by Bowdler (1977) almost 40 years ago, but still relevant to this debate. That observation raises the question of the delay we might except between the initial colonization of the then exposed continental shelf and the appearance of people in distant locations: in the south or east, and in the desertic inland. We cannot exclude the possibility of extremely rapid settlement of all regions, perhaps by highly mobile foragers moving rapidly across landscapes using a narrow dietary focus, which might have minimised the gap between colonization and occupation of the sites mentioned here (see Beaton 1991). Such a rapid dispersal rate would be congruent with the calculations by Birdsell (1957) and is supported by a number of researchers including O'Connell and Allen (2012). Alternatively, in any region, it might have taken some time for population levels to grow to a point where material indicators of human activity became sufficiently common that they have been preserved and can still be found by archaeologists today, and hence humans may have been present in landscapes for a prolonged period before they become archaeologically visible at about 50,000 years ago. This proposition does not suggest any particular dispersal rate, but it implies that the current sample of dated assemblages across Australia is not so much an indication of the initial appearance of humans in each region as it is a record of demographic growth and residency length at each location. This proposition is applicable to not only dated artifact assemblages, but also to the onset of higher charcoal deposition rates in sediment cores. While those charcoal spikes were initially interpreted as indicators of the arrival of people, they are increasingly interpreted as indicating the emergence of altered landscape management strategies after the extinction of mega-marsupials, perhaps substantially after people arrive in that landscape.
Human Contributions to Transformations in Australia's Flora and Fauna
The nature and extent of human impact on the Australian landscape is a subject of great contention, and, like debates in the Americas, arguments have polarized into those in favor of severe and rapid humanly induced changes and those that minimise the contribution of humans to shifts in fauna and flora communities. The environmental transformations most discussed include the reduction and extinction of megafaunal browsers and grazers across the continent. A suite of very large animals had lived in Australia at some time prior to the arrival of humans: giant kangaroos (such as Macropus rufus and Macropus giganteus titan) and giant wombat (Phascolonus gigas), tall flightless birds (Genyornis sp.), and four-legged marsupial browsers and grazers the same size as some species of hippopotamus and rhinoceros (such as Diprotodon optatum, Zygomaturus sp., Palorchestes sp.). Some studies have claimed a broad coincidence between the time at which species of large marsupials disappeared and the time that humans arrived (Miller et al. 1999; Roberts et al. 2001), and have argued that the human colonization of Australia might have triggered a trophic collapse in which particular kinds of animals were driven to extinction. As skillful predators whose hunting behaviors were unfamiliar to marsupial prey, the dispersing humans no doubt had the capacity to reduce the viability of vulnerable species, especially if hunters targeted the young of animals that reproduced slowly or if the prey species were limited in distribution and predictable in movement by, for example, being tethered to rare resources (Brook and Bowman 2002, 2004).
Excerpted from Paleoamerican Odyssey by Kelly E. Graf, Caroline V. Ketron, Michael R. Waters. Copyright © 2014 Kelly E. Graf, Caroline V. Ketron, and Michael R. Waters. Excerpted by permission of Texas A&M University Press.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.