Scholars of Pleistocene human antiquity have long debated both the timing and the route by which humans first arrived in the New World. Until recently, the most widely accepted explanation centered on the Clovis hunter-gatherer culture as the initial inhabitants. The Clovis people are thought to have migrated to the Americas during the last Ice Age by traversing a land bridge between northeastern Asia and North America.
Over the last decade, the Clovis-first model has been challenged by an accumulation of evidence from sites that predate the arrival of the Clovis culture in North America. In their paper, “Late Upper Paleolithic Occupation at Cooper’s Ferry, Idaho, USA, ~16,000 Years Ago,” Loren Davis et al. use radiocarbon dating to establish a pre-Clovis chronology for artifacts and evidence of human activity from one of these sites.
The Cooper’s Ferry site in Idaho is located at what was once the southern margins of a vast continental ice sheet, known as the Cordilleran, that covered northwestern North America at different points during the Late Pleistocene. From the latest discoveries at Cooper’s Ferry described in their paper, Davis et al. find support for the notion of a dispersal route along the Pacific coast for the first human inhabitants of the New World. They also suggest the possibility of a technological link between northeastern Asia and a pre-Clovis population in the Americas.
The artifacts most commonly associated with the Clovis culture are large fluted projectile points. These points have been found at numerous sites throughout North America and are notable for the consistency of their shaping and design. The stemmed projectile points found at Cooper’s Ferry were an early clue that the site might predate the Clovis culture. From their design it was clear that they belonged to an entirely separate technological group, later termed the Western Stemmed Point tradition.1 In their paper, Davis et al. provide updated results from dating efforts on the oldest evidence of human occupation at the site.2 They provide radiocarbon dates for charcoals and faunal remains correlated to specific stratigraphic sequences, along with data gathered from lithic artifacts, namely stemmed projectile points.
As part of their study, Davis et al. report a dating of between ~15,660 and 13,260 cal. BP (calibrated years before the present) obtained from LU3, a stratigraphic unit comprised of sedimentary layers formed from an accumulation of wind-blown silt. These dates, corresponding to a period during the Late Glacial era (16,000–13,000 cal. BP), are considerably older than the Clovis complex,3 currently dated to ca. 13,300–12,800 cal. BP.4 Four pits that were excavated at the same level as LU3 also yielded pre-Clovis dates, ranging between ca. 15,000 and 13,000 cal. BP. From Bayesian analysis and modeling, Davis et al. found that the Cooper’s Ferry site was first occupied as early as 16,560–15,280 cal. BP.5 This date range is thought to precede the appearance of ice-free overland routes across northwestern North America during deglaciation. Davis et al. theorize that the earliest occupants of the Cooper’s Ferry site most likely migrated along the northern Pacific coast.
Morphological similarities between the stemmed projectile points found at Cooper’s Ferry and bifacial stemmed points found in Japan are the basis for a cultural connection that Davis et al. suggest may have existed between the pre-Clovis people and those of the Upper Paleolithic in northeastern Asia.6 This proposal, it should be noted, is in line with the latest genetic findings that Late Pleistocene human migrations from East Asia contributed to the ancestral populations of Native Americans.7
One of the most significant implications of the oldest dates obtained at Cooper’s Ferry is that the site is now formally recognized as being a pre-Clovis occupation. Other recognized pre-Clovis occupations in North America include the Manis site near Sequim in Washington, the Paisley Caves in Oregon, and the Gault and Friedkin sites in Texas.8 These sites are all comparable in age to the Monte Verde II site found in a coastal region of southern Chile and dated at ca. 14,800–13,500 cal. BP.9 This chronology fits with current estimates for the dispersal speed of the first humans to reach the southern end of South America.
Genomic data obtained from the remains of thirty-four ancient individuals found in northeastern Siberia has offered tantalizing clues about the origins of the first inhabitants of North America.10 The remains date from the Pleistocene and Holocene (31,600–600 cal. BP) and suggest that the ancestors of Native Americans were diverged from the Late Pleistocene populations in Siberia. In North America, ancient DNA from the Anzick infant, the remains of a Paleo-Indian male found in south-central Montana, confirms that the Clovis people contributed to the Native American population.11 It remains unclear whether the pre-Clovis peoples coexisted with any other genetic populations in northeastern Asia during the Late Pleistocene. Where these pre-Clovis populations may have originated is also unknown. According to current hypotheses, the northeastern Asian genetic population that contributed to the dispersal to eastern Beringia,12 and subsequently North America, were likely present in southern Beringia during the Last Glacial Maximum, ca. 26,000–19,000 years ago.13
Although Davis et al. do not attempt to identify a genetic ancestor for the pre-Clovis—such as the Ancient Paleo-Siberians and East Asian populations identified by Martin Sikora et al. in another study published in 201914—their paper suggests that the Upper Paleolithic population of the Japanese archipelago was a likely candidate. With the exception of the Ryukyu Islands in southwestern Japan,15 the Pleistocene human paleontological record in Japan is extremely limited and the genetic structure of the Paleolithic population remains shrouded in mystery. In the absence of direct evidence, archaeological inferences play an important role in the study of pre-Clovis origins. The challenge then becomes to establish a valid inference involving a human dispersal event from the Japanese archipelago to other areas in northeastern Asia, such as Beringia, that could have preceded a coastal migration to the New World.16
As part of their study, Davis et al. compare stemmed points from the oldest occupation (LU3) at Cooper’s Ferry to Japanese bifacial stemmed points, notably the Tachikawa-type stemmed points (TSP) found in northern Japan. The comparison is based on morphological similarities and a common dating to the Late Glacial era (16,000–13,000 cal. BP). While the overall shapes of both types of stemmed point do exhibit some common features, a great deal of caution should be exercised when attempting to link technologies with human dispersal. This is especially true for comparisons involving such distant locations on either side of the Pacific.
The TSP designation was originally assigned to projectile points excavated from Localities II and III of the Tachikawa site in southern Hokkaido. The first study of these points was published in 1960.17 Three distinguishing characteristics were identified: a bifacial projectile point shaped by pressure-flaking techniques; a stemmed portion that is about a quarter of the overall length of the point; and a stem base with grounded lateral edges.18 Subsequent discoveries blurred the boundaries of this relatively straightforward classification. These included other stemmed points found in Hokkaido, such as the styles known as Shukubai and Engaru. The stemmed points classified as TSP are now subdivided into five classes based on morphological traits in the proportions and shape of their stemmed bases.19
The time frames during which TSP were manufactured remain unclear. The artifacts at the original Tachikawa site were shallowly buried, and chronometric evidence was not available.20 Even though a number of stemmed-point types from Hokkaido—such as the Tachikawa, Engaru, and Shukubai types—have been assigned to the late Upper Paleolithic,21 their precise chronometric ages are still not known.22 By contrast, stemmed points found on Honshu have been firmly dated to ca. 14,000 cal. BP. This time frame is also associated with the production of a style of earthenware pottery thought to originate in the earlier Jōmon period of Japanese prehistory. Known as Ryukisenmon, the pottery dates from ca. 15,000 to 14,000 cal. BP and is attributed to the Late Glacial interstadial period.23 Jōmon, which can be translated as “rope-patterned,” and Ryukisenmon refer to a style of linear-relief pottery decorated using the impression of thinly coiled rope and smoothed with a spatula. The earliest Jōmon pottery found on Hokkaido is the Tsumegatamon, or nail-impressed ware, from the Taisho 3 site on the eastern side of the island, firmly dated using radiocarbon techniques to 15,000–14,000 cal. BP.24
When attempting to explain similarities in the form and style of artifacts between different sites, it is necessary to distinguish between homology, convergence, and cultural drift.25 The human dispersal from northeastern Asia to North America was a unique event in human history. As a result, archaeologists have tended to focus more on similarities than differences in their units of observation—e.g., classes of stone tools, types of stone material, and groups of common features. The people of the Late Pleistocene were modern humans who would have possessed behavioral plasticity.26 As they dispersed along the circumpolar northern Pacific and throughout Beringia, changes in the periglacial environment they encountered would have necessitated periodic reorganizing of their subsistence behavior and technology. These changes could equally have resulted in technological convergence or variability.
In 2003, I attended at a graduate seminar taught by Bruce Huckell and Joseph Powell at the University of New Mexico. This seminar was my first exposure to Paleo-Indian research and I was immediately struck by the many controversies and open questions surrounding the origins of the Clovis and pre-Clovis peoples. These included the mysterious finds from Sandia Cave in New Mexico,27 first discovered during the mid-1930s, and contested hypotheses attributing European origins to the Clovis people, such as the Solutrean hypothesis.28
Since the mid-1990s, the work of American researchers examining the origins of the Clovis people had been complicated by the discovery of Kennewick Man. Found by a pair of college students on a riverbank in Washington State, the skeletal remains were identified as those of a Paleo-Indian male. His morphology suggested a connection to the Ainu indigenous population in northern Japan. The discovery was followed by a decades-long legal battle to establish ownership of the remains. Researchers fought for control of the skeletal remains and artifacts against both the US Army Corps of Engineers, who were responsible for the area in Washington State where the remains were found, and local indigenous groups, who claimed the remains under the Native American Graves and Protection Act.29 The issue was not fully resolved until 2017 when the remains were returned to the local Umatilla people and reburied.
While the Kennewick Man controversy was unfolding, the prospect that coastal migration was the most likely route by which early humans arrived in the Americas began gathering support.30 Over the same period, new research reshaped a perception of Paleo-Indians as primarily a cultural entity associated with the population groups that gave rise to the ancestors of Native Americans. The nature of the pre-Clovis debate changed as a result. In particular, phylogenetic relations among Paleo-Indian complexes have been investigated using geoarchaeological approaches, which address archaeological questions using methods from the earth sciences, along with quantitative analyses focused on stone tools.31 In their paper, Davis et al. examine the northeastern Asian origins of the first dispersed population in the New World using evidence gleaned from the latest advances in genetic research. As the title of their paper suggests, the notion of the Paleolithic in American prehistory should be expanded to include the late Pleistocene technological connection to the Old-World Upper Paleolithic.32
