Introduction: an archaeological signature of our arrival?
One challenge to the human evolutionary sciences is to relate the overall trend of Pleistocene human morphological evolution to changes in human lifeways. The simplest picture is that these lifeways changed, as individual humans evolved slower life histories and enhanced cognitive and physical capacities. Changes in individual capacity opened the door to new technical and social skills, and hence to new lifeways. This picture predicts a direct relationship between evolutionary changes in human morphology (reflected in speciation) and changes in the archaeological record. This direct connection is obscured to some degree by signal loss, and by economic, ecological and demographic constraints on lifeways. However, over large temporal and spatial scales, environmental factors are expected to vary enough to allow potential capacities to be expressed, at least in some times and places (Tennie et al. 2017 Reference Tennie, Braun, Premo, McPherron, Haidle, Conard and Bolus2016; Tomasello Reference Tomasello2020).
Intrinsic individual capacity maps directly onto lifeways when social learning is not cumulative; that is, when social learning only enables people to acquire capacities they could acquire individually. In such cases, in Claudio Tennie’s terminology, lifeways are constrained within their ‘zone of latent solutions’ (see, for example, Tennie et al. Reference Tennie, Premo, Braun and McPherron2017). When lifeways are thus constrained, marked changes reflect changes in that zone. To the extent that social learning is cumulative, ecological, social and technical capacities expand beyond that zone. The crucial aspect of cumulative social learning is that it enables individuals to acquire information or skill they could not acquire by individual experimental learning. However, it is still often supposed that archaeological signals of cumulative culture themselves reflect a major upgrade in intrinsic capacity, with the idea that cumulative social learning depends on distinctive cognitive adaptations (Gamble et al. Reference Gamble, Dunbar and Gowlett2014; Tomasello Reference Tomasello2020). Our paper argues instead that cumulative social learning emerges incrementally, dependent on a mosaic of social and cognitive factors. As human evolution proceeded, the archaeological record became increasingly the product not just of intrinsic individual capacity, but social organization, cumulative cultural learning and individual adaptive plasticity. Human cognitive capacities are in part shaped by previous histories of cultural accumulation, not just genetic inheritance (Colagè & d’Errico Reference Colagè and d’Errico2020; d’Errico et al. Reference d’Errico, Doyon and Colage2018). Consequently, the record becomes increasingly decoupled from intrinsic individual potential. The ethnographic record documents great diversity in social organization and lifeway, without difference in the genetic potentials of these varying peoples. We suggest that this decoupling of lifeway and innate capacity has deep historical roots. We develop this picture through a discussion of the Middle Stone Age (MSA) of Africa and behavioural modernity.
The African MSA saw significant lifeway changes roughly coincident with the emergence of sapiens as an identifiable taxon. As Kandel and colleagues have shown, the MSA record shows great regional and temporal variation, perhaps with some signal of adding innovation over time (Kandel, Bolus et al. Reference Kandel, Bolus and Bretzke2016). The African MSA began about 300 kya, fading out about 30 kya. MSA lithic technology is characterised by a near-absence of the Large Cutting Tools, and by the central role of prepared core technology (Kandel et al. Reference Kandel, Bolus and Bretzke2016). The latter requires extensive pre-shaping of a core, including a much-modified striking platform, which in turn allows flakes of a roughly predictable form to be detached, each with a single strike. Prepared core technologies are arguably more transformative and more complex than the technologies of the Acheulean, but support a far wider range of tool morphologies (Muller et al. Reference Muller, Clarkson and Shipton2017). Discoid techniques also make it easier to correct errors through further core preparation. Thus one advantage of prepared core techniques was simplification of the technical demands of toolmaking, for with Acheulean technology, every strike matters, and error correction is more restricted (Liu et al. Reference Liu, Khreisheh, Stout and Pargeter2023; Muller et al. Reference Muller, Sharon and Grosman2025).
As the MSA developed, other contrasts with the Acheulean emerged, with more long-distance raw material transport (whether by exchange or increased mobility); more fire use (Thompson et al. Reference Thompson, Wright and Ivory2021); more frequent hafting and more use of bone for identifiable uses (points to haft, for example) (Kandel et al. Reference Kandel, Bolus and Bretzke2016; Scerri & Will Reference Scerri and Will2023). There is some evidence of broader ecological tolerances (Roberts & Stewart Reference Roberts and Stewart2018), with humans systematically exploiting coastal resources and establishing in rainforests and deserts, though as always, some caution is needed here given the effects of temporal bias on these records. As with prepared core technologies, these other contrasts are typically taken to be signs of a larger and more complex social life, of an increase in individual cognitive capacity, or both. This is especially true of a MSA development that many theorists take to be especially significant: a record of an increasing use of material symbols (Wadley Reference Wadley2001; Reference Wadley2015).
In describing the MSA, it is important to emphasize the fact that the archaeological record is not simple and unilinear. A basic set of MSA competences seems to have established early and spread widely across Africa: prepared core technology, hafting and the use of stone from relatively distant sources (Scerri & Will Reference Scerri and Will2023). Other innovations seem to be later, many between 200 and 100 kya (Stewart & Jones Reference Stewart, Jones, Stewart and Jones2016), highly regional. These persist over enough space and time to be archaeologically visible, but nonetheless disappear from the record. The pace of innovation and regional differentiation increases further after about 100 kya (Wadley Reference Wadley2021), but still with the same pattern of regionally distinctive material cultures establishing, persisting for some thousands of years, then disappearing. For example, at the southern African site of Vasche River there is evidence both of very early ostrich-egg flasks and of a distinctive form of heat treatment, making possible the production of small cores. This complex persisted from about 90 to 80 kya, and then the record becomes akin to other sites in the same general fynbos region (Mackay et al. Reference Mackay, Armitage and Niespolo2022).
The earliest fossils claimed as sapiens are north African, at about 305 kya (Hublin et al. Reference Hublin, Ben-Ncer, Bailey and Freidline2017), though these do not have the characteristic sapiens globular skull. Indeed, the full suit of characters regarded as sapiens markers (enhanced encephalization, globular skull, small gracile face, chin) are not found in the record for perhaps another 200,000 years (Meneganzin et al. Reference Meneganzin, Pievani and Manzi2021). Despite these complexities, it has been suggested that the MSA is the archaeological signature of the evolution of the sapiens mind, perhaps at the origins of the MSA, and making this transition possible, or perhaps during the MSA, helping to explain further developments (Roberts & Stewart Reference Roberts and Stewart2018; Wadley Reference Wadley2015). On this second view, the extended and complex evolution of the full morphological suite of characters distinctive of sapiens in conjunction with the increase in complexity of technology and social life through the MSA reflects the incremental evolution of an increasingly modern-like mind (Kandel et al. Reference Kandel, Bolus and Bretzke2016; Roberts & Stewart Reference Roberts and Stewart2018).
We are sceptical of all models of the MSA that see it as essentially an expression of newly evolved intrinsic cognitive capacities. We shall argue in the next section, ‘Technology, economy and the social world’, that cumulative social learning does not depend on a specific evolutionary innovation enhancing social learning, nor is there a cognitive or social threshold dividing those taxa capable of cumulative social learning from those that are not. Here we are in agreement with Meneganzin and Currie (Reference Meneganzin and Currie2022), though our analysis depends on a specific hypothesis about cumulative culture rather than general evolutionary considerations. As cumulative culture depends (we argue) on a variable suite of cognitive, social, demographic and economic factors, its archaeological signal is of slow establishment, spread and stabilization. This is simultaneously a signal of the slow emancipation of technical, social and economic capacities from the intrinsic cognitive capacities of the agents in question. Human lives are less and less expressions of Tennie’s ‘zone of latent solutions’.
It is true that human evolution from about the late Pliocene has seen an encephalization trend: later humans tend to be larger, with relatively larger brains. Moreover, there must be some causal relationship between increasing cognitive capacity and change in technology. It is almost certain that the Heidelbergensians, living around 800 kya, had skills out of the reach of end-Pliocene habilenes. However, over deep time, human social lives, and the archaeological traces of those lives, have become less tied to the intrinsic traits of individuals. Showing this is the aim of the next section. Section 3, ‘Did MSA technology require enhanced cognitive capacities?’, applies that analysis to the MSA, arguing that its most distinctive features may well be the result of social and environmental changes, changes that are themselves significantly decoupled from changes in individual capacity. We suggest that the MSA is in part a signature of the increasing autonomy of human social worlds from innate individual capacity: they are not, even to a first approximation, a simple reflection of the intrinsic capacities of the agents who make them up. Section 4, ‘Building behavioural modernity’, takes up behavioural modernity in the light of this analysis, presenting a revisionary account of its nature and causes. An analysis of behavioural modernity identifying the causes as primarily social and economic, and the effects as primarily an expansion of variation, inevitably raises a question about our sister taxon, the Neanderthals, and we discuss the implications of this view for Neanderthals as we conclude. Sterelny (Reference Sterelny2011) also saw behavioural modernity as the result of stable cumulative cultural learning, but without appreciating the richness of earlier phases of the MSA, the importance of the forager education system or the extent and importance of neural plasticity.
Technology, economy and the social world
Current orthodoxy suggests that, while the social environment is important, its main characteristics are fixed by intrinsic individual capacities. Social worlds change in response to changes in individual capacities. In particular, intrinsic individual capacities are seen as (i) the primary determinants of the size and complexity of social groups (see for example Gamble et al. Reference Gamble, Dunbar and Gowlett2014; Gowlett et al. Reference Gowlett, Gamble and Dunbar2012); (ii) the primary determinant of the efficiency of social learning (see Pradhan et al. Reference Pradhan, Tennie and van Schaik2012; Tennie et al.2020) Reference Tennie, Premo, Braun and McPherron2017; ; (iii) the primary determinant of social lives mediated by symbols and hence by a conscious, explicit awareness of individual and social identity (as in Klein Reference Klein2008; Tattersall Reference Tattersall and Gonthier2021). On this view, larger and more connected social worlds, with efficient, high-fidelity social learning, and with social life organized around an explicit understanding of who we are and how we differ from others, are the essential ingredients of recognizably human social lives. But these features of social life were made possible by, and are a consequence of, the biological evolution of intrinsic individual capacities.
That orthodox analysis should be rejected. While human social worlds are markedly influenced by individual capacity, they are not a simple reflection of individual capacity. Causal influence flows both ways. As individual capacity increased over evolutionary time, making possible richer and more complex social worlds, the reciprocal influence of social world on individual capacity increased. The lives of those who live in contemporary industrial societies illustrate these reciprocal effects. Many such people are literate, numerate and have immediate and reliable access to vast amounts of information, all acquired over generations through feedback between individuals and the social scaffolding in which they are embedded. Their cognitive powers—capacities to think, not just what is thought about—depend on being embedded in these worlds (Clark Reference Clark2008). These effects of social environment on individual capacity have deep histories. As Lyn Kelly has documented, non-literate cultures use many technologies that support memory and navigation (Kelly Reference Kelly2015). We discuss this interaction below, through three uncontroversially central phenomena: social scale, social learning and the use of material symbols.
Social scale
Robin Dunbar, with various colleagues, compared social taxa with non-social relatives. On his analysis, the data supported both a qualitative claim that social scale correlates with relative brain size (or perhaps relative neocortex size), and a stronger, quantitative claim: a relationship between relative neocortical volume and maximum stable group size. This gave a ‘Dunbar number’ for a given species.Footnote 1 These empirical observations are supported theoretically by the observation that when social interaction is mediated by mutual knowledge, memory demands rise faster than group size, as relationships matter, and the number of relationships in a group much exceeds the number of individuals in a group. Dunbar then retrofits his analysis to both sapiens and earlier humans (see for example Dunbar Reference Dunbar and Ingold1994; Reference Dunbar1998; Reference Dunbar2003; Gamble et al. Reference Gamble, Dunbar and Gowlett2014). On his analysis, sapiens individuals can keep rough track of something like 800–1000 individuals, but have much smaller networks of those they know well: around 120. Those numbers are smaller for earlier humans (Gamble et al. Reference Gamble, Dunbar and Gowlett2014).
Social scale and connection matter. Larger and better-connected groups generate, preserve and share innovation much more reliably than smaller and more fragmented groups (Premo & Kuhn Reference Premo and Kuhn2010). Reliable cumulative culture probably required social networks of significant scale and richness; perhaps even as rich as those documented ethnographically. For example, Kim Hill and his colleagues have shown that a typical forager, over the course of his/her life, will have seen hundreds of others making tools (Hill et al. Reference Hill, Wood, Baggio, Hurtado and Boyd2014). So if social network size is tightly controlled by encephalization, social networks with the scale and richness sufficient to support reliable cumulative culture, as evidenced in the MSA, may have required sapiens-scale encephalization. However, we doubt that network size is controlled by encephalization (or by any other intrinsic feature of neural architecture). For one thing, as human social worlds from the Neolithic onwards show, social networks can be managed and maintained by mutual recognition of social role (Seabright Reference Seabright2010). At most, social networks managed by mutual personal knowledge are constrained in the way Dunbar proposes. But even that is very doubtful. Memory and the ability to map one’s social world can be enhanced by cognitive tools and external scaffolds. Many now use diaries, photograph albums, Facebook, and the like; social networks do not have to be managed by neural resources alone in keeping track of individuals and their relationships (Clark Reference Clark2008). As we noted earlier, this is not just a feature of WEIRD cultures (Western Educated Industrial Rich Democratic) (Kelly Reference Kelly2015). Memory is scaffolded by names, stories, dress, styles of personal adornment, modes of interaction and living sites. The role of kinship is less central in the large-scale societies of the industrial west, but even there a personal name typically codes genealogical and geographical information, as does common membership of clubs and similar institutions. In more traditional societies with greater emphasis on kin connection and less idiosyncratic naming practices, more is coded more reliably. Moreover, individuals are not solitary actors when maintaining and updating their social maps. In interaction, we remind and prompt one another, as John Sutton and collaborators have shown (Michaelian & Sutton Reference Michaelian and Sutton2019; Sutton et al. Reference Sutton, Harris, Kell and Barnier2010).
In sum: neither social scale nor social organization are tightly constrained by agents’ innate cognitive powers.
Social learning
Social learning is widespread across the tree of life, but cumulative social learning is rare amongst nonhuman animals. In contrast, in the human lineage, cumulative social learning has played an important and increasing role through the Pleistocene. Michael Tomasello, and more recently Claudio Tennie, with various colleagues, have argued that cumulative cultural learning requires high-fidelity social learning, and that in turn depends on a specific, late-evolving cognitive adaptation, imitation learning (and perhaps also an advanced form of theory of mind) (see for example Tomasello Reference Tomasello1999a,Reference Tomasellob; Reference Tomasello2020; and Henrich & Tennie Reference Henrich, Tennie, Muller, Wrangham and Pilbream2017; Tennie et al. Reference Tennie, Braun, Premo, McPherron, Haidle, Conard and Bolus2016; Reference Tennie, Premo, Braun and McPherron2017; Reference Tennie, Hopper, van Schaik, Hopper and Ross2020). Imitation is a routine sapiens accomplishment, often part of play, but it requires the novice to transform a visual representation of an operation into a motor sequence representation, perhaps via an intermediate egocentric visual representation. This cognitive challenge demands a specific adaptation, a fact that explains why cumulative culture is so rare outside our species. The contrast in the pace of change and extent of regional differentiation between the Acheulean and the MSA allegedly indicates that Acheulean humans had at most rudimentary adaptations for social learning (Tennie et al. Reference Tennie, Call and Tomasello2009).
Learning by imitation is indeed important for some social skills: for example, dance and ritual (Jagiello et al. Reference Jagiello, Heyes and Whitehouse2022). But this analysis grossly exaggerates its importance. Imitation plays little role in the acquisition of many crucial forager skills. For:
i. High-fidelity imitation enables an agent to learn a precise motor routine by observing a model using that routine, as in learning a dance step. But forager skills often do not involve specific motor routines: consider tracking, learning local geography, natural history, fire management. These are all complex skills, almost certainly built over many generations of cumulative improvement (Boyd Reference Boyd2016). But they are not motor procedures, or bundles of motor procedures (Lowe Reference Lowe2002). That is true even of fire management: much of the skill consists in the identification of suitable fuels, in siting and building fireplaces so that they draw rather than smoke, in the control of heat and in the many uses of fire (Garde et al. Reference Garde, Nadjamerrek, Kolkkiwarra, Russell-Smith, Whitehead and Cooke2009). While becoming an expert tracker requires social support and mentoring (Liebenberg Reference Liebenberg1990; Shaw-Williams Reference Shaw-Williams2014), tracking is not a motor skill at all. Fire ignition depends primarily on well-designed fire bows or fire drills; the physical skill is simple repetitive movement (Blake & Welch Reference Blake and Welch2006).
ii. Other forms of forager expertise do require finely honed motor skills, but even these are rarely learned by imitation (for ethnographic examples of blade work and trap making, see Boyette Reference Boyette2016; Naveh Reference Naveh, Terashima and Hewlett2016). Natural materials are often too variable for an artisan to succeed only by learning and applying a fixed procedure. Heidelbergensians could not craft handaxes by reproducing a pre-learned mechanical procedure (as one might assemble a kitset chair by following the illustrations of one being assembled). The substrates (and task demands) vary too much: cobbles vary indefinitely in their initial size and shape, and in their propensities to fracture. Stout and colleagues have analysed the production sequences required to make Acheulean and Oldowan tools, but the atomic units of their sequences are not motor acts but subgoals (‘prepare striking platform’) (Stout Reference Stout2011; Stout et al. Reference Stout, Hecht, Khreisheh, Bradley and Chaminade2015). Likewise, György Gergely and Idikó Király have reanalysed the original experimental work on children’s ‘over-imitation’, showing that strictly speaking the children were not imitating motor patterns, but identifying and reproducing subgoals. Instead of bowing towards a touch-sensitive globe, they switched the light on by lifting it to their head, touching it that way (Gergely & Király Reference Gergely, Király and Charbonneau2024). One form of social learning is learning by observation, but often that learning proceeds by understanding the structure of the task, rather than by recalling and reproducing a particular motor pattern (Morin Reference Morin2016; Sterelny & Hiscock Reference Sterelny and Hiscock2024). Having identified a subgoal, the novice’s motor pattern may match that of the model, but it need not, as Gergely and Király show, and often novices have to adapt a motor procedure, given differences in dexterity, grip strength and handedness. Tools made of natural materials, and made by hand, are less standardized, and standardization is necessary to apply a fixed routine (Eerkens & Bettinger Reference Eerkens and Bettinger2001). Only industrial technology enables products, parts and tools to be more standardized, and hence enables assembly to be reduced to pre-specified stereotyped procedures. Moreover, there is significant debate as to whether imitation learning really does require cognitively demanding capacities. Cecilia Heyes has shown that at least some forms of imitation can be learned through simple associative mechanisms. In her view, sapiens imitation evolved through changes in motivation, not cognition (Heyes Reference Heyes2018; Reference Heyes2021).
As the ethnography of forager childhoods shows, young foragers reliably acquire the informational capital of their community (Lew-Levy et al. Reference Lew-Levy, Reckin, Noa, Cristóbal-Azkarate and Ellis-Davies2017; Reference Lew-Levy, Kissler, Boyette, Crittenden, Mabulla and Hewlett2020). We return to this ethnography in section 4, ‘Building behavioural modernity’. Here it is enough to note that children learn by doing, but with rich social support and social inputs. They have plenty of time to play and explore, with much redundancy and learner control over the pace and targets of learning, so this education system depends on efficient adult foraging, generating sufficient surplus to afford children the opportunity to build the competences they will need (Kaplan et al. Reference Kaplan, Hooper and Gurven2009). In this system, information arrives through many natural and social channels. The crucial abilities are those of integrating information from multiple sources, assessing one’s own performance and using information from these many channels to improve that performance (often iteratively) (Sterelny Reference Sterelny2021; Reference Sterelny2023). There is little doubt that these capacities have improved over evolutionary time. While it is likely that intrinsic changes in neural size and organization have played some role in this dynamic, there is no cognitive threshold, no cognitive divide between agents capable of participating in cumulative culture and those that are not. The possibility of cumulation will depend on complex interplays between learning targets, life history, individual cognitive capacity, social mechanisms that mitigate information loss and social supports for learning.
Material symbols
Human social groups and their members are aware of themselves as distinct and identifiable, and they signal this awareness of identity and difference. Clothes, styles of personal adornment and the organization of living space often express individual and social identity (Hewlett et al. Reference Hewlett, Hudson, Boyette, Fouts, Lavi and Friesem2019). This is a fundamental feature of human sociality. Ethnographically known foragers live not just in residential groups nested in larger social wholes, they consciously identify as members of specific communities, with this identification often reinforced by a distinctive language or dialect; shared origin myths; distinctive rituals (often enhanced by material symbols); sometimes distinctive styles of dress and adornment; and norms of social interaction. This shared understanding links networks of residential groups, making it possible for the groups to aggregate, and for individuals and families to move within the network (Bird et al. Reference Bird, Bird, Codding and Zeanah2019). Sometimes it makes larger-scale collective action possible (Boyd & Richerson Reference Boyd and Richerson2022). Many technologies of material symbol production are archaeologically fragile, possibly masking even earlier origins of these practices (Stibbard-Hawkes Reference Stibbard-Hawkes2025), but evidence of material symbols mediating this shared understanding dates from the MSA. This includes the production of pigments, personal adornment, incised ochre, ultimately rock art.
It is often suggested that the organization of social life through conventional and arbitrary symbols requires all the resources of the modern mind (for instances, see Davidson Reference Davidson2010; Gamble et al. Reference Gamble, Dunbar and Gowlett2014; Tattersall Reference Tattersall2016; von Petzinger Reference von Petzinger2017; Wadley Reference Wadley2001; Reference Wadley2015). For example, Gamble, Dunbar and Gowett argue that material symbols can be produced and understood only by agents with sophisticated theories of mind. Symbols are produced with complex communicative intentions, and are understood only if those intentions are recognized. An ochre pattern is painted on a shield with the expectation that this design will be understood as an intent to communicate, and with the further expectation that the audience will thereby recognize the belief the agent intends to communicate: this waterhole is his. Perhaps some material symbols require such sophistication, but many do not. Their significance can be learned by association. When, say, a distinctive ochre pattern on a shield denotes membership of a group, these insignias of identity co-occur with those thus identified (McDonald & Harper Reference McDonald and Harper2016). The significance of these symbols can be learned by simple association.
Rather than material symbols being the result of cognitive change in the human lineage, we suggest that they became more visible in the MSA as the social world became more complex and more connected. This explanation fits better with the existence of some material signals in the pre-MSA record. Some handaxes seem to have been made for display rather than use (Kohn & Mithen Reference Kohn and Mithen2009; Nowell & Chang Reference Nowell and Chang2009).Footnote 2 It also fits the patchy record of material symbols in the MSA record, as size and connectivity are likely to have varied markedly over space and time. There is little call for signals of identity or role when communities are small, relatively homogenous and fairly isolated (Kuhn & Stiner Reference Kuhn, Stiner, Mellars, Boyle, Bar-Yosef and Stringer2007a,Reference Kuhn and Stinerb). But once residential groups became networked into the nested multi-level structures known from ethnography, salient, readily recognizable signals of individual and corporate identity became important, often integrated with stereotyped, sometimes very dramatic action sequences. In their ethnography of Central Australian foragers at the end of the nineteenth century, Baldwin Spencer and Frank Gillen published many stunning photos of these dance-material symbol hybrids. These include rituals of approach: when parties of semi-strangers (known but not intimates) arrive at an aggregation site for ritual business, they must signal their identity and peaceful intent (Spencer & Gillen Reference Spencer and Gillen1899). Jerome Lewis has documented music-ritual-material symbol hybrids from Central Africa with similar social roles (Lewis Reference Lewis, Botha and Knight2009; Reference Lewis2015).
These considerations suggest that (a) the size and connectivity of the social world, the extent and nature of support for social learning, and individuals’ time and energy budgets are all relevant to the retention of informational capital and the rate of innovation; (b) these factors are influenced but not determined by the intrinsic capacities of the agents in question; (c) hence they are likely to vary over space and time, independently of the taxonomic identity of those agents. The archaeological novelties of the MSA may well be no more than a consequence of the social environment becoming somewhat more favourable to retention and innovation, independently of the taxonomic identity of the personnel in those social networks.
In sum: social scale, the efficiency of intergenerational social learning, and mediating social interaction through signals of social identity and role (rather than direct personal knowledge) all shape human lifeways. But they are not direct reflections of individual capacity, and there is little reason to suppose that changes in their salience over the MSA are a result of the biological evolution of new capacities.
Did MSA technology require enhanced cognitive capacities?
The MSA record is one of impressive technical, social and ecological change. As noted in the first section, MSA humans (i) regularly made stone tools using prepared core techniques. (ii) Hafting became a common feature of MSA lifeways, implying as well soft material technologies of wood working, cordage and adhesives (Barham Reference Barham2013). (iii) The heat treatment of stone appears, first at Pinnacle Point, about 160 kya, and is common during the subsequent MSA (Schmidt & Högberg Reference Schmidt and Högberg2018). (iv) The use of bone and allied materials to make identifiable tools (like points) becomes an established part of human technical competence (Wadley Reference Wadley2015). (v) Evidence of high-velocity projectile weapons (perhaps launched javelins) appears, perhaps as early as 100 kya (O’Driscoll & Thompson Reference O’Driscoll and Thompson2018). (vi) While domesticated fire preceded the MSA, fire becomes regularly used, indicating reliable fire-lighting capacities (Gowlett Reference Gowlett2016). (vii) Niche breadth expanded: convincing evidence of systematic exploitation of coast resources; suggestive but less certain evidence of the occupation of rainforest and semi-arid and arid habitats (Roberts & Stewart Reference Roberts and Stewart2018). (viii) the record shows the appearance and expansion of material symbols. Many of these innovations appear quite late in the MSA, and even after their appearance, most have an episodic, patchy record. Even so, it is an impressive array of innovation (Kandel et al. Reference Kandel, Bolus and Bretzke2016; Scerri & Will Reference Scerri and Will2023; Stewart & Jones Reference Stewart, Jones, Stewart and Jones2016).
The highly encephalized trio of sapiens-Neanderthals-Denisovans surely had cognitive capacities beyond those of their ancestors, and of course it is possible that those enhanced capacities supported this history of innovation. That said, once social learning is sufficiently reliable and with sufficient bandwidth to sustain a broad skill set and a significant technical repertoire, accelerating innovation is no surprise. Richer technical repertoires scaffold their own expansion via positive feedback effects. Modular technologies offer opportunities for recombination; techniques invented for one tool (an adhesive for hafting) can be adapted for other purposes. Fire, too, is productive, as heat treatment of stone shows, but it also eases time budgets, increasing the value of food and lowering the cost of tinkering, by creating more usable hours in the day (Wrangham Reference Wrangham2017; Wrangham & Carmody Reference Wrangham and Carmody2010). Moreover, the more technology there is, the more information about technology is stored in the world. Not all tools can act as templates for their own reproduction, but many can, and all are proof of their own possibility: seeing a hafted knife shows that hafting is possible. All technology is informational technology. Finally, increased encephalization combined with morphological constraints on infant brain size required a greater proportion of brain maturation to take place after birth and under varying environmental influence, and hence resulted in greater phenotypic plasticity (Zerilli Reference Zerilli2018; Reference Zerilli2019). Cognitive capacities are consequences as well as causes of social lives and social learning (Heyes Reference Heyes2018).
These considerations suggest that the impressive record of innovations later in the MSA may have been the result of positive feedback loops initiated by the early establishment and wide geographic spread of those base technologies. While it is possible that intrinsic changes in neural architecture were essential to the establishment of these MSA technologies and the feedback loops they supported, there is suggestive evidence that those base technologies were not outside the cognitive reach of pre-MSA humans. For these base technologies all have pre-MSA antecedents: prepared core techniques, hafting, points and microliths, the use of bone are all known from the pre-MSA archaeological record (Brooks et al. Reference Brooks, Yellen and Potts2018; Wilkins & Chazan Reference Wilkins and Chazan2012; Wilkins et al. Reference Wilkins, Schoville, Brown and Chazan2012). It is true that these pre-MSA records are rare: the techniques seem not to have spread in time and space. Perhaps these skills were at the limits of cognitive possibility for pre-MSA humans. If so, that would explain why these MSA techniques did not become standard features of the lifeways of the large-brained humans that lived before sapiens and its sister taxa. and why they were not then available as platforms for further diversification and improvement. The core MSA techniques became an engine of change only for later humans with even greater cognitive capacity. This is certainly possible. But there are reasons for caution; the more parsimonious alternative cannot be excluded.
First, it is probable that biases in the archaeological record make these pre-MSA traces of MSA technology seem even rarer than they really were. The older the site, the less likely it is to survive. Additionally, if there were fewer pre-MSA humans, there would have been fewer and/or smaller assemblages to begin with. To some extent, the fainter signal of these earlier, MSA-like technical capacities, compared to the somewhat stronger MSA signal, reflects these biases. More importantly, there is an alternative, economic explanation of why these technologies did not become characteristic of pre-MSA sites, even though the precursors to sapiens were capable of their discovery and use. In general, cost is inversely related to frequency. In most sites, the most frequently encountered tools are simple unretouched flakes made from local material (for an Australian example, see Moore Reference Moore2013). As they are more expensive, more complex tools will be a smaller fraction, perhaps a much smaller fraction, of total production. Moreover, it is likely that the costs of hafting, microliths, and prepared core tools were relatively greater for pre-MSA humans, further depressing their relative frequency. Brian Hayden (Reference Hayden1998) has drawn archaeological attention to a general feature of the life history of technical innovation. When they first appear, innovations are often prestige items (hafted tools, microliths, obsidian points, the first pots, the first metal tools). They are expensive: made only with exotic or highly processed materials (metal, obsidian, heat-treated silcrete) and/or because only a few can make them well. Within individual history of skill acquisition (as in driving a car), there is a well-known transition from reliance on conscious, explicit, slow, attention-demanding guidance (‘type two’ cognition) to fluent, rapid, routinized, automatic guidance (‘type one’ cognition) (Christensen et al. Reference Christensen, Sutton and McIlwain2016; Reference Christensen, Sutton and Bicknell2019). In many cases, there is a similar transition in the life history of a skill. A skill that is initially difficult, mastered by only a few, and requiring care and attention by that few, becomes part of the standard repertoire of the community. Typing illustrates this trajectory: in the nineteenth century, a highly paid prestigious male occupation; by the mid-twentieth, a low-prestige, low-paid female occupation. Technologies which once signalled prestige, like owning a pot or a metal knife, become standard items of kit. This transition is in part the result of optimizing and simplifying the skill itself, accelerating production and lowering opportunity costs, and in part finding better ways of scaffolding skill transmission. These are standard features in the life history of technical skills.
The implication of this analysis is that in their initial uses, hafted tools, microliths and flakes made from prepared cores were even more expensive than they became once skill acquisition was better supported by competent adults, with skills becoming typical within the community. At the early stage of invention/adoption, when such tools are made only by a few, they were rarely made, and the incoming generation had fewer opportunities to learn by observation and emulation. Since the new tools were rare and expensive, children would have had fewer opportunities to learn by experiment and play. Forager ethnography suggests that children usually have a lot of freedom to experiment with adult equipment not in use, but that is unlikely to be affordable for expensive items of kit. As a consequence, new skills were vulnerable to demographic attrition. When only a few individuals in a community have a skill, it can be easily lost through migration or a few unlucky deaths. This risk of loss would be further exacerbated if pre-MSA communities tended to be smaller and/or more isolated. Isolation means that lost skills are less apt to be regained from adjacent communities, and in smaller communities, with fewer skilled models, novices have to invest more heavily to acquire more challenging skills, reducing the incentive for their acquisition (Sterelny Reference Sterelny2020). These economic and demographic factors impede uptake and spread of innovations depending on complex operations and long learning histories. While it is possible that hafting and the control of flake form by core preparation were too cognitively challenging to become a widely established feature of pre-MSA lifeways, these considerations provide an alternative explanation.
Moreover, there is independent evidence of cumulative social learning by pre-MSA humans. On some views, this dates back to the very beginning of flake production (Paige & Perreault Reference Paige and Perreault2024). While this may be an overcall, some cumulative social learning has deep roots. A distinctive feature of Pleistocene humans was their ecological adaptability: they were able to occupy and exploit a wide range of habitats. If the dating of a Chinese site is correct, hominids first expanded out of Africa by 2 mya (e.g. Lordkipanidze et al. Reference Lordkipanidze, Ponce de León and Margvelashvili2013; Zhu et al. Reference Zhu, Dennell and Huang2018). While their Eurasian habitat preferences might somewhat have resembled East Africa, there were always significant differences in flora, fauna, climate, seasonality and landscape structure. These early migrants established in areas different both from one another and from their African homelands. This required adapting to novel and changing circumstances. Moreover, there is persuasive evidence of large game hunting (or at least confrontational scavenging) prior to 1 mya. With short-range weapons, this must have depended on skill, knowledge and cooperation (Pickering Reference Pickering2013). Sapiens exemplify this adaptability particularly strongly, probably being the first to establish in truly arid lands, high latitudes and true rainforest. But exploiting any new habitat requires communities to learn about their resources and dangers and to develop appropriate skills to exploit the first and manage the second (Henrich Reference Henrich2016). Since forager lifeways depend on rich but concealed and/or defended resources, those lifeways are knowledge and skill intensive. Acquiring this informational capital is a multi-generation process, as evidenced by the sad fate of many European explorers dying (or requiring local rescue) through their inability to acquire essential information (Richerson & Boyd Reference Richerson and Boyd2005).
In summary: if pre-MSA foragers could not accumulate cultural information, they would have been excluded from much of their out-of-Africa range. Pre-MSA cumulative culture is shown by Acheulean technology itself; by at least some control of fire; by the natural-history skills and organization capacity for profitable hunting and confrontational scavenging. While the scope of accumulation was limited, as seen in the transient character of further technical innovation, these limits may well have been consequences of demographic and economic factors above. Likewise, early migrations out of Africa into many habitats show some capacity to accumulate adaptive information. While it is true that, especially after about 100 kya, sapiens lived in a broader range of niches than their predecessors, it is much less clear that they did so because they were sapiens. While it is possible, perhaps even likely, that the record of technical innovation in the MSA depended to some degree on both changes in intrinsic neural architecture and the positive feedback factors identified above, neither the initial emergence of the MSA, nor developments within in it, are clear signals of sapiens speciation or of enhanced innate cognitive capacities. The null hypothesis cannot be rejected: social, demographic and environmental factors suffice to explain these changes independently of changes in intrinsic neural architecture or taxonomic identity. As the technical repertoire becomes richer, as the social world increasingly scaffolds social learning, innovation and differentiation increase. Moreover, as encephalization enhances adaptive phenotypic plasticity, cognitive capacity itself becomes a result as well as a cause of human lifeways.
Building behavioural modernity
We see behavioural modernity as entering the range of variation of mobile forager cultures documented in ethnography. This variation is a contingent moment in time, in a world before the expansion of sedentary life and farming. But it is a contingency with a deep history of multiple independent and semi-independent trajectories of cultural evolution, and in widely different physical and biological habitats. Common features are likely to be robust rather than accidents of time and place. Though ethnographic information has serious limits, we still have quite rich information about these mobile forager social worlds and the economic and social factors that support them. The archaeological record has some information about when Pleistocene communities began to fall into that range of variation and, to some degree, about the genealogy of the factors that support this variation.
As noted earlier, temporal and economic biases in the record might well have erased the earliest evidence of this expanded variation. Despite these biases, by MIS (Marine Isotope Stage) 4, there is evidence of technical sophistication, some use of material symbols and artifacts made from non-local materials. But at least by MIS 3 or MIS 2 archaeological evidence suggests that sapiens communities fell within the range of variation known from ethnography, perhaps with the exception of an ability to carry through large-scale collective action projects. To date, evidence for large-scale fish weirs and traps, extensive and expensive game-drive lines, and boat construction requiring hundreds of skilled person days, all derive from the Holocene (see for example Boyd & Richerson Reference Boyd and Richerson2022; Frison Reference Frison2004; McNiven et al. Reference McNiven, Crouch, Richards, Sniderman, Dolby and Corporation2015). The records of MIS 3 and MIS 2 show: (i) utilitarian technologies as diverse and sophisticated as those documented ethnographically: hunting weapons included bows, javelins, harpoons, knives; domestic equipment included hafted axes but also awls and other tools for the preparation of skins and fitted clothing; obvious evidence of full control of fire, including its use as an ecological tool; indirect evidence of water transport, and direct evidence of exploitation of marine resources via debris from pelagic fish; systematic evidence of shelters; (ii) rock art, ochre, items of personal adornment, in a few places figurines and music instruments, evidence of intense but non-utilitarian activity in secluded places. These are all signs that ritual, or something like ritual, played an important role in these Late Pleistocene communities; (iii) many sites in these periods clearly show the use of non-local materials, transported or traded from their place of origin, most likely showing that residential groups were not sealed in place and isolated from others; (iv) by MIS 2 (but probably earlier), foragers had established in the high latitudes, arid lands, rainforest, and perhaps high altitudes; (v) perhaps most important of all, there is evidence of sustained differentiation: of self-sustaining cultural-technological blocks (Kuhn Reference Kuhn2020; Wadley Reference Wadley2021).
When these have contingent elements, as in different choices of material substrate for material symbols, but also in different choices about foraging tools, they suggest durable,Footnote 3 spatially extended cultural networks, distinct from others through social tools rather than physical barriers. In turn, this suggests life in multi-layered, self-aware communities known, in many different forms, from forager ethnography (Layton et al. Reference Layton, O’Hara and Bilsborough2012). These differentiating and self-sustaining blocks are major engines of forager variation in the ethnographic present, though of course differentiating responses to environmental variation are likewise important. The Australian record from MIS 3 provides a clear example, with hafted, edge-ground axes a signature tool of northern Australia, but not found in southern sites of the same age, despite their being no overt barrier to north–south movement. There are African examples, including those that significantly predate MIS 3, and these might well indicate the earlier establishment of impressively varied and stably distinct lifeways. We have already noted the distinctive, locally restricted but temporally durable technical traditions of Varsche River. The Still Bay technical suite of southern Africa persisted from around 75 to 70 kya. This repertoire included red pigment use, shell beads, bone carving, heat treatment and pressure flaking. Somewhat later, this cultural block was succeeded by the Howiesons Poort repertoire. Rather than marine shell, beads were made from ostrich eggs, and projectiles were tipped by small standardized arrowheads rather than heat-treated and pressure-flaked bifacial spearheads (Henshilwood & Dubreil Reference Henshilwood and Dubreil2011). Ceri Shipton has documented and discussed a similar sequence of locally stable, somewhat durable technical repertoires at the Kenyan site of Panga ya Saidi, a site with a continuing occupation sequence over the last 78,000 years (Shipton Reference Shipton2024).
The analysis of this paper suggests that feedback between at least four factors made possible this combination of adaptation to specific ecological challenges, broad ecological tolerance, technical-social sophistication and extended social networks.
Demography
In the literature, the idea that behavioural modernity was the result of an increase in social scale has been the main alternative to the view that it resulted from a leap in intrinsic cognitive capacity. The central ideas of this line of reasoning are: (i) the archaeological signs of an increase in the rate of innovation and of the extent of regional diversity substantially postdate the origins of sapiens and are episodic in time and variable in space, so the decisive change cannot be a change in intrinsic cognitive capacity (though this might be a necessary background condition); (ii) instead, the crucial factor is change in the scale and connectivity of sapiens populations (see (Powell et al. Reference Powell, Shennan and Thomas2009). While demography matters, the most critical change is enhanced connectivity. In itself, an increase in population density (perhaps as a side-effect of concentration in refugia, as might have occurred on the coasts of South Africa in arid/cold periods), might well impede cultural accumulation. Great ape residential groups are closed, with few positive interactions between residential groups. Had human groups been closed in the same way, an increase in population density would just intensify competitive interactions. Sapiens social worlds are open, both in the sense that there is often quite free movement between residential groups in connected networks, and in the sense that in most forager cultures, residential groups aggregate from time to time. This shift to more open networks was probably both incremental and gradual, but the stable cultural blocks identified above indicate that open networks were in place in the Late Pleistocene. However, as open social worlds were managed by new social tools, open networks and cumulative culture coevolved. Ethnography suggests that open networks are maintained through elaborated kinship systems (made possible by elaborate kinship terminologies); material exchange; common or overlapping ritual identity; (often) a distinctive language or dialect (Chapais Reference Chapais2008; Reference Chapais2013; Layton & O’Hara Reference Layton, O’Hara, Dunbar, Gamble and Gowlett2010; Layton et al. Reference Layton, O’Hara and Bilsborough2012; Sterelny Reference Sterelny2019). These social tools all forge, or make more salient, affiliative relations between agents in different residential groups. As they are the products of cumulative culture, the relationship between enhanced connectivity and more reliable cultural accumulation is not cause and effect, but co-evolution. The social tools that support enhanced connectivity were built through cumulative social learning.
The forager education system
Ethnographically described forager cultures depend on the forager education system. Forager education turns out to have a remarkably consistent cross-cultural pattern. It is characterized by the following broad tendencies: (i) a forager’s education develops over three main phases: toddlers, older children, and adolescents. Toddlers learn primarily from their parents (mostly the mother). From childhood to early adolescence (from about 4 years old to about 12), the young forager spends most of the day in a mixed-age/mixed-gender play-group. Within this group, learning is a mix of collaboration, individual exploration, practice, horizontal transmission, but with some information flow through participation in adult activities. Their learning is scaffolded by both miniaturized equipment made for them, access to adult gear and freedom to use it. They primarily learn from doing, but with rich social support from peers and adults. From early adolescence, social learning becomes predominantly intergenerational, with a variable mix of oblique and vertical transmission, often from close kin. (ii) Forager children are free-range. The play-group operates with remarkably little adult supervision, with the children having considerable control over their own time budgets, giving them time to play and explore. Yet those children seem eager to acquire adult skills and incorporate them in further play/exploration. (iii) Forager children acquire a basic forager competence remarkably early. By their early teens they are still to acquire only their community’s most demanding skills. (iv) Elaborated, explicit teaching is mostly restricted to norm acquisition, esoteric knowledge and the most challenging subsistence and technological skills. Explicit teaching is mostly unobtrusive and low cost. (v) Children are included in much of the profane adult world. Forager campsites are compact and intimate, with much of domestic life in the open (Hewlett et al. Reference Hewlett, Hudson, Boyette, Fouts, Lavi and Friesem2019). The intimacy of these living situations allows children to look, to eavesdrop, to learn by helping and to benefit from occasional advice. This framework is remarkably effective in building the competences needed for adult life (for reviews of this ethnography, see Hewlett & Lamb Reference Hewlett and Lamb2005; Hewlett et al. Reference Hewlett, Fouts, Boyette and Hewlett2011; Konner Reference Konner, Hewlett and Lamb2005; Lew-Levy et al. Reference Lew-Levy, Reckin, Noa, Cristóbal-Azkarate and Ellis-Davies2017; Reference Lew-Levy, Lavi, Reckin, Cristóbal-Azkarate and Ellis-Davies2018).
This organization of education is impressively adaptive, combining efficient transmission with a reduction of the impact of children on their parents’ time budgets and, to some degree, their resource budgets. Early competence makes those young partially self-supporting. However, this system itself had preconditions which themselves evolved over considerable time, probably through some form of gene–culture coevolution. It requires adult impulse control and exceptional tolerance of the young (Wrangham Reference Wrangham2018). In addition, forager education required high levels of adult foraging efficiency. As we shall see, this might be an important difference between Late Pleistocene sapiens and Neanderthals. This efficiency makes possible a slow life history and a long period of scaffolded learning that reciprocally facilitates cumulative culture (Kaplan et al. Reference Kaplan, Gangestad, Gurven, Lancaster, Mueller, Robson and Roebroeks2007; Reference Kaplan, Hooper and Gurven2009; Koster et al. Reference Koster, McElreath and Hill2020). Moreover, this efficiency buys the resources needed for innovation. Established experts are likely to be the source of incremental improvements in existing technologies: for example, improving point designs. But innovations by recombination might be found in play and exploration. It may well be that the use of cordage in snares, in tripwires and bird traps originated in play. A relaxed time and energy budget also matters for adult innovation, especially if there were transition costs to a new technology. Consider, for example, the transition from spear/javelin hunting to bow hunting. There is a substantial fitness trench between these hunting modes. Only a well-made, powerful bow, well-engineered to avoid component failure, with someone skilled in its use, offers a superior rate of return to hunting with a javelin and spear-thrower. Moreover, given the challenges of bow-making, a good deal of experiment and practice would be required to go from a low-powered, dubiously accurate bow to one with heavy draw weight matched with arrows suited to give real hitting power. This was not an innovation transition available to a community under serious resource stress. As well as efficient provisioning, the forager system also required foragers to dominate the carnivore niche. Free-range childhoods with little adult oversight would be much too dangerous unless predators had largely learned to avoid foragers and their campsites. More generally, slow sapiens life histories depend on low levels of extrinsic mortality, including low levels of deaths from predation (Guthrie Reference Guthrie and Roebroeks2007).
Technology breeds technology
A third factor is the self-accelerating effect of the growth of technology and its skill base. As noted in section 2 above, modular multi-component tools open up new regions of design space. Techniques invented for one tool (an adhesive for hafting) can be adapted for other purposes (Kuhn Reference Kuhn2020). Once a community can make and haft microliths, new options are only an innovation or two away: multi-pronged fish-spears; fletched projectiles, harpoons, shafts with multiple armatures; scythes, and saw-like tools for processing soft materials but with the safety and mechanical advantage of a handle. Richer technical repertoires scaffold their own expansion via positive feedback effects. Cordage is an especially productive technology. Once a community can make twine, their technical possibilities expand enormously: shelters; rafts and other boats; fitted clothing; nets; snares, traps; harpoons; bows; an array of musical instruments. More generally, as the base skill set expands and improves, potential innovations come within reach. MSA lithic technology shows this kind of expanded design space as knappers altered the scale and geometric form of prepared cores, to diversify flake formation form. Finally, the more technology there is, the more information about technology is stored in the world. Many tools can act as templates for their own reproduction. Technology is a cognitive scaffold.
Teaching old brains new tricks
The cognitive foundation of cumulative social learning was the ability to integrate information across a range of social and natural sources. Integration is central to adult forager lifeways, too. In adaptive decision making, cooperative foraging requires an agent to combine information about social partners (their skills, information, motivation, physical strength) with information about their natural environment (spatial geometry, risks, location, abundance and net value of different resources (see, for example, the discussion of fire drives in Garde et al. Reference Garde, Nadjamerrek, Kolkkiwarra, Russell-Smith, Whitehead and Cooke2009). These abilities to integrate and weigh information from different sources were needed by pre-sapiens humans, and there is no sharp threshold that makes accumulation possible. We presume these capacities have increased over time, with changes in neural architecture playing some role. But increased plasticity may be as important. Cecilia Heyes has recently argued that cognitive capacities that enhance social learning—most obviously language, but also imitation and theory of mind—are themselves learned socially, with these cognitive tools built by cumulative cultural learning (Heyes Reference Heyes2018). She points to literacy (and numeracy) as proof of concept. Writing systems evolved incrementally and culturally, and these external representational systems evolved and became important parts of agents’ environments. While not everyone masters these systems, fully literate minds adapt to use them fluently and effortlessly. Heyes’ expansive claims are controversial, but not her more modest claim that social learning influences how we think, not just what we think about (Heyes Reference Heyes2012). Ivan Colagè and Francesco d’Errico have applied a similar picture to the late MSA archaeological record (see Colagè & d’Errico Reference Colagè and d’Errico2020; d’Errico & Colagè Reference d’Errico and Colagè2018; d’Errico et al. Reference d’Errico, Doyon and Colage2018).
Conclusion
This paper has focused on the African MSA as a critical period of behavioural evolution prior to significant geographical expansions of Homo sapiens. We have argued that behavioural modernity, which we take to be the development of the variation in lifeways found in ethnographic foragers, is not primarily explained by the emergence of sapiens as a species. It is largely a consequence of informationally driven cultural accumulation. We take this to have more general import: the continuation of a deep time trend in human evolution of the gradual decoupling of human lifeways from the intrinsic capacities of individuals. The amplification of this decoupling trend is fundamental to understanding the accelerated changes and increased regional and temporal variation from the start of the MSA.
The extent to which Neanderthals followed a similar trajectory of further technical innovation and increased ecological breadth for similar reasons remains unclear. The trend in Neanderthal archaeology is of a diminishing gap between Neanderthals and contemporaneous sapiens, with most of the differences quantitative (Shea Reference Shea2023, chapter 8), and with Neanderthal sites showing some features not found in sapiens sites (Stiner & Kuhn forthcoming). If the analysis of this paper is correct, we would expect at most a muted version in the Neanderthals of the sapiens MSA trajectory. There is suggestive evidence that Neanderthal social worlds were smaller and less well connected than those of sapiens (Churchill Reference Churchill2014). If so, the innovation rate would be depressed relative to sapiens, and those innovations would be more vulnerable to demographic attrition. In addition, it is probable that Neanderthal foraging did not generate the surplus akin to that funding the forager system. There is evidence that Neanderthal children faced resource stress (Shea Reference Shea2023, chapter 8). Neanderthals needed more calories per capita, and there is suggestive evidence that the impressively efficient forager division of labour was not part of their lifeways (Stiner & Kuhn forthcoming), perhaps because that division could only establish in subtropical habitats, even though once established it could be modified for others (O’Connell Reference O’Connell and Conard2006). Lifeways without much surplus would both make intergenerational learning less reliable and communities less able to pay transition costs to incorporate new techniques into their lifeways. Collectively, these differences from African sapiens of the MSA would limit potentials for innovation and technical expansion and make Neanderthal residential groups especially vulnerable to fragmentation and isolation.
Despite its non-linear and regionally variable character, the MSA record is indeed one of considerable social, technical and ecological innovation. Towards MIS 3 or before, the result was human lifeways falling within the range of variation known from ethnography. These developments reflect a piecemeal establishment of more reliable cumulative social learning, to which genetically evolved changes in the human lineage may have made some contribution. In our view, the most important cognitive change was an increased ability to integrate and assess information arriving through a range of social and natural channels. But we have additionally argued that the MSA record has also been profoundly shaped by social, demographic and economic features of MSA residential groups. One consequence of increased phenotypic plasticity in the human lineage, especially of behavioural and cognitive plasticity, was an increase in the effects of varied and changing social environments on those agents. Cognitive capacities became effects as well as causes of changes in human lifeways and technical repertoires. Biological changes in capacity acted in concert with behavioural factors: the richness of the existing technical repertoire; the organization of childhood and social learning; at times, at least, a relaxed time and energy budget allowing adults to experiment and innovate; increased connectivity and cooperation between residential groups. Human lifeways at the end of the Pleistocene were the result of positive feedback amongst these factors, rather than being the upshot of a single revolutionary transition between MSA times and those of earlier humans.
Open access
