We thank our colleagues for their insightful comments. The weight of modern evidence is against the notion that contemporary human cultures can be tracked backwards into the Pleistocene (e.g. Lee & DeVore Reference Lee and DeVore1976; Kuper Reference Kuper1988; Wilmsen Reference Wilmsen1989; Solway & Lee Reference Solway and Lee1990; MacEachern Reference MacEachern2000). Modern-day hunter-gatherers are not our Stone Age ancestors. Current protestations notwithstanding, the provocative title that d'Errico and colleagues (Reference d'Errico, Backwell, Villa, Degano, Lucejko, Bamford, Higham, Colombini and Beaumont2012) chose for their paper, ‘Early evidence of San material culture represented by organic artifacts from Border Cave, South Africa’, unambiguously asserts the opposite. Our critique of that paper's content does not question the robusticity of the methods employed at Border Cave (for this, see Evans Reference Evans2012). Rather, our comments focus on the theoretically flawed search for a specifically ‘San’ “cultural adaptation” (d'Errico et al. 2012: 13214) at any Pleistocene archaeological site.

Historically, southern African archaeologists have used ‘San’ ethnography to frame Later Stone Age archaeology (e.g. Goodwin & van Riet Lowe Reference Goodwin and van Riet Lowe1929), but this practice has more recently been roundly criticised (e.g. Parkington Reference Parkington1989; Mitchell Reference Mitchell, Mitchell, Haour and Hobart2003; Humphreys Reference Humphreys2005; Sealy Reference Sealy2006). The societies lumped together as ‘San’ reflect millennia of interaction, contact and acculturation, resulting in their diverse material cultures and worldviews. Barnard's (above) encounter with a Kalahari man who spoke five distinct languages highlights the degree to which the term ‘San’ oversimplifies contemporary cultural variation. As Denbow notes above, extracting so-called ‘San’ cultural elements from the “contexts of change and variation in which they were embedded” greatly reduces their variability and complexity. In ignoring this variability/complexity, we allow ethnography to dictate higher-order theory in archaeology. Our alleged ‘re-hashing’ of this point in several venues is a measure of a problem that d'Errico and colleagues dismiss.

D'Errico et al.’s (2012) choice of examples for “early evidence of San material culture” at Border Cave, and their failure to comprehend our Table 1, highlights the variability and complexity that has been overlooked. For example, they lump the Border Cave bone ‘tools’ into morpho-functional categories whose integrity remains untested. Do we know that the pointed objects at Border Cave were used as weapon tips rather than as needles, awls or any of the other myriad functions to which a pointed bone object might have been used? Plainly, we do not. The fact that both the Netsilik Inuit and some ‘San’ groups used bone tools does not prove relatedness, but it does point to broader patterning in this class of object. We partially agree with d'Errico and colleagues that such classes are “too broad” to be a “serious means of demonstrating or disproving anything”. Consistency in coarse classes does not prove continuity. Where coarse classes are absent from large numbers of assemblages, we cannot simply draw a dotted line through time and choose to believe that they are masked by taphonomy or recovery techniques.

Discoveries of bone tools or ostrich-eggshell ornaments at sites beyond southern Africa (e.g. Kabwe, Mumba and Magubike) suggest patterning outside the regions typically associated with ‘San’ ethnography (see Wadley Reference Wadley2015 and references therein). Their discovery at sites pre-dating the Later Stone Age (e.g. Sibudu Cave and Klasies River Mouth), at sites excavated before the “careful excavations conducted at Blombos, Diepkloof and Sibudu caves” (d'Errico et al. above) (including Boomplaas, Cave of Hearths and Bushman Rockshelter), and their absence from recently excavated sites with excellent preservation (e.g. Klein et al. Reference Klein, Avery, Cruz-Uribe, Halkett, Parkington, Steele, Volman and Yates2004) surely attests to the fact that their absence in various parts of the southern African archaeological record is more real than apparent. The fickle nature of these remains is further accentuated upon removing the Still Bay ostrich-eggshell ornaments from our Table 3 (their original inclusion in our manuscript was an error, and one inconsequential to our argument; we thank d'Errico and colleagues for noting this error, although they were unwilling for it to be corrected in our paper as published). We observe, however, the presence of notched bone in both the Still Bay and post-Howiesons Poort techno-complexes at Apollo 11 and Sibudu Cave (Cain 2004; Vogelsang et al. Reference Vogelsang, Richter, Jacobs, Eichhorn, Linseele and Roberts2010), the purpose of which is currently unknown. This pattern cannot simply be ascribed to poor organic preservation, inadequate recovery techniques or the size of excavations. As Lombard and Parsons (Reference Lombard and Parsons2011) note, variation of this kind in the late Pleistocene of southern Africa now forms one of the most exciting topics in the region's hunter-gatherer prehistory.

D'Errico and colleagues (above) rhetorically ask, “What do we know about the possible resilience of cultural traits in hunter-gatherer populations living in southern Africa between 40 kya and the present?”; they then answer, “Virtually nothing”. Actually, we know quite a bit (e.g. Deacon Reference Deacon1984; Parkington Reference Parkington1984; Mitchell Reference Mitchell1988; Wadley Reference Wadley1993; Sealy Reference Sealy2006). We do not know as much as we might wish, but we know enough to reject the argument that ‘San culture’ remained recognisably the same for 40 millennia.

Finally, it is encouraging to note that d'Errico and colleagues have begun more systematic ethnoarchaeological research among Kalahari populations, and that Border Cave is being reinvestigated. These more behavioural, ecological and comparative approaches, grounded in sounder understandings of both ethnographic and archaeological contexts, are surely better ways to move forward (e.g. Shea Reference Shea2011). We join Wobst (above) in hoping that this work leads to the investigation of more “anthropologically interesting relationships or processes” than can be attained from a misplaced search for ‘origins’ or primordial ‘Pleistocene San’ cultures.