Klein (1989, 1994, 1999, 2009), Stringer and Gamble (1993), Mellars (1996), Berwick and Chomsky (2016), Tattersall (2017), and others have speculated that a recent mutation related to neural circuitry was responsible for the emergence of modern behavior. (Although Gamble no longer subscribes to the theory.) The mutation is posited to be recent in the sense that it is assumed to have been under selection in African H. sapiens populations after our last common ancestor with other hominin; in particular Neanderthals and Denisovans. The phenotypic trait under selection is assumed to relate to the wiring of the brain since it cannot be identified morphologically. In what follows we shall argue that the recent mutation hypothesis cannot be reconciled with the available body of evidence.

Chomsky’s fundamental insight was that the language capacity was confined to our species and uniform across it. To wit, if you raise a Japanese child in Bangladesh, she will speak Bengali like a native. This implied that under all the massive synchronic (cross-sectional) and diachronic (temporal) variation in languages lies a simple structure for which Chomsky coined the term Universal Grammar. Building on that foundation and a deeper appreciation of the role of stochastic factors in evolution, Berwick and Chomsky (2016) argue that the posited mutation endowed the possessor with an innate capacity for generative grammar (2016, p. 87):

At some time in the very recent past, apparently some time before 80,000 years ago if we can judge from associated symbolic proxies, individuals in a small group of hominids in East Africa underwent a minor biological change that provided [the basic structure of generative grammar] … The innovation had obvious advantages and took over the small group. … In the course of these events, the human capacity took shape, yielding a good part of our “moral and intellectual nature,” in Wallace’s phrase. The outcomes appear to be highly diverse, but they have an essential unity, reflecting the fact that humans are in fundamental respects identical, just as the hypothetical extraterrestrial scientist we conjured up earlier might conclude that there is only one language with minor dialectal variations, primarily—perhaps entirely—in mode of externalization.

The precise mechanism posited by Berwick and Chomsky (2016) sounds plausible. But, as we shall show, it cannot be sustained in light of the entire body of evidence. Even if a mutation was responsible for the human capacity for language, we will argue, it cannot explain the emergence of modern behavior. The weight of the evidence suggests that, if such a mutation indeed occurred, it did so much earlier and was demonstrably shared by the Neanderthals and presumably other big-brained hominin during the Upper Pleistocene, 200-30 Ka.

There are three kinds of evidence of relevance to the question at hand: fossil remains, archaeology, and DNA. Briefly, fossil remains in the form of bone fragments, teeth, postcranial skeletons allow us to reconstruct the morphology of Pleistocene hominins. Hominin taxa (human “species”) are defined with reference to standard taxonomic rules on the basis of morphological criteria (more precisely, autapomorphies). The precise details need not concern us beyond the fact that hominin taxa, H. sapiens, H. neanderthalensis and so on, are defined by morphological traits, not DNA.

The archaeological evidence consists of artifact assemblages that capture the material culture of Pleistocene hominin populations. Modern behavior, in as much as it can be operationalized, has to be pitted against this evidence. The clearest evidence of modern behavior in Pleistocene assemblages is symbolic storage—personal ornaments, decorated tools, engraved bones and stones, grave goods, formal use of space, and perhaps style in lithics. We’ll return to that evidence presently. But note that is a micro-local definition. Modern behavior can also be recognized in the overall dynamism suggested by rapid cultural turnover in assemblage artifacts. Indeed, the very idea of modern behavior comes from the contrast between later dynamism and the extraordinary monotony of Oldowan and Acheulean lithic assemblages during the first two million years after the rise of the genus Homo. Oldowan stone tools from c. 2.5 Ma show very little variation over a million years even as an adaptive radiation sends H. erectus to expand into Asia (‘Out of Africa I’). Virtually identical Acheulean handaxes were manufactured for another million, 1.5-0.5 Ma, throughout the western Old World. (It is speculated that bamboo tools replaced handaxes east of the Movius Line.)

All of this begins to change in the Upper Pleistocene, 200-30 Ka (we prefer the more neutral climatic label modified by the stratigraphic term ‘Upper’ for ‘Late’ which serves to clarify the archeological criteria being used to define the period). Lithic assemblages suddenly begin to display turnover in style and regional variation. The complexity of artifactual assemblages starts to accelerate. With the Upper Paleolithic in western Eurasia at the end of the Upper Pleistocene, we get the “full package” of modern behavior described by Diamond as “the Great Leap Forward.” It is this (eurocentric) body of evidence that suggests a decisive structural break in behavioral patterns that the neural mutation hypothesis is mobilized to explain. If true, the hypothesis would indeed explain this broad diachronic pattern in Pleistocene assemblages in the western Old World. But as we shall see, the hypothesis runs into insurmountable difficulties when broader diachronic and synchronic patterns are examined with a finer brush.

Great hopes were once placed on DNA. It was assumed that the problem of identifying the molecular basis of many if not most phenotypic traits would be solved sooner or later. All such hopes have since been dashed. The problem has turned out to be much more complex than previously imagined. What DNA has turned out to be good for is identification (DNA as molecular fingerprinting), uncovering phylogenetic relationships (who is closer to whom ancestrally), and importantly for us—with DNA recovered from ancient fossils—population history.

It is extremely difficult to make kosher inferences about Pleistocene history from the DNA of contemporary populations because of what Lahr (2016) calls ‘the Holocene Filter’ 15-0 Ka. Human population exploded a thousandfold from a few million c. 15 Ka to a few billion today. Genetic evolution accelerated to 100 times the rate that had prevailed prior to the Neolithic. And massive ‘population pulses’ (such as the massive migrations that followed the Neolithic c. 9 Ka and Secondary Products Revolution c. 5 Ka) transformed the population structure of macroregions, replacing ancient paleodemes of Pleistocene hunter-gatherers by mixed populations, with later arrivals as the predominant element (eg, the Neolithic and Yamnaya pulses in western Eurasia and the Bantu expansion in southern Africa). For all these reasons, modern populations cannot be identified with ancient populations; not locally, not regionally, and not globally. Pleistocene hominin populations, including anatomically modern humans, were very different from contemporary populations in both phenotypic traits and DNA. This means in particular that the frequency of modern-archaic admixture cannot be reliably ascertained from the DNA of living populations. We must rely instead on ancient DNA.

We cannot be sure if other taxa in the genus Homo had the capacity for generative grammar. But if we are to run a horse race in behavioral traits between “moderns” and “archaics”, we must have a level playing field. The period of extraordinary monotony in lithic assemblages is one of exponential (“hockey-stick”) encephalization. Aiello and Wheeler (1995) have shown that the brain and the gut are the heaviest consumers of metabolic energy; so that for brains to grow, guts had to shrink. This required a shift to higher quality foods; in particular, carnivory (first through scavenging and later through hunting; although true hunting did not emerge until much later). Moreover, it was accompanied by bigger bodies with greater metabolic demand. Bigger brains and bodies in turn required more complex (and need we say, successful) hunting and foraging behavior.

Dunbar (1998)’s social brain hypothesis tied the neocortex ratio of mammals to the size of their social network, suggesting that it was the computational demands of social life that selected for encephalization. At any rate, the powerful feedback loop generated runaway encephalization culminating in the Upper Pleistocene with big-brained (~1400 cc), big-bodied hominin like ourselves.

So the genus Homo is too big for the purpose. It includes taxa with brains not much larger than chimps as well as taxa with brains bigger than ours. Clearly, we must compare H. sapiens with other encephalized hominin. We therefore introduce an Encephalization Filter that rules out all but the late archaic hominin. See next figure.

What all this means is that in order to compare apples to apples, we must restrict the test universe to Upper Pleistocene Homo. Otherwise we might as well compare modern day New Yorkers to Neanderthals and conclude that the latter are nonmodern in their behavior (and, by implication, subhuman tout court). The acid test of the recent mutation hypothesis is whether we exhibited more modern behavior than the others in our period of overlap. In sum, the relevant test universe consists of clusters of paleo-demes (prehistoric geographically-situated populations) from multiple taxa in the genus Homo during the Upper Pleistocene, 200-30 Ka.

Ultimately, whether we call these taxa species or subspecies is irrelevant—they are defined by their autapomorphies. Still, Wolpoff and Caspari wonder if the Upper Pleistocene taxa now called “species” don’t satisfy the textbook definition of subspecies. Indeed Mayr has shown that the average splitting time for speciation is in millions of years. According to Henneberg (1995), the doyen of hominin taxonomy, ‘until thoroughly falsified, the Single Hominid Lineage Hypothesis provides less complicated description of hominid evolution … It is also compatible with the uniformitarian postulate — the recent human evolution occurred, and occurs, within one lineage consisting of widely dispersed but interacting populations.’ These frames are quite consistent with each other.

Within the single human lineage that is called the genus Homo there has always been a bushy tree with population structure. There is no reason to believe that paleo-demes were completely genetically isolated from each other. Demes exchange genes at the margins of ecoregions which is how gene flow is maintained within a geographically dispersed species. The point is that under Pleistocene levels of population density and the attendant isolation-by-distance, there obtained intercontinental population structure (ie continental races or discrete subspecific variants) of the sort that has no modern counterpart. Simply put, continental populations during the Upper Pleistocene were considerably further away from each other in neutral phenotypic and genotypic distance than they are today or were in the ethnographic present. Wolpoff and Caspari lost their big battle over the model for the ethnographic present in favor of the ‘(Mostly) Out of Africa’ model. But their multiregional model works nicely for Upper Pleistocene population structure. As Zilhão notes in the European context, ‘it is highly unlikely that the Neandertal-sapiens split involved differentiation at the biospecies level.’

Hominin taxa now called “species” are better conceptualized as geographic clusters of paleo-demes situated in macroregions under relatively severe isolation and therefore derivation; maybe not enough derivation for speciation but perhaps enough for raciation (400-800 Ka, Cf. the San splitting time from the rest of the extant human race ~280-160 Ka).

During the Upper Pleistocene, Neanderthals were endemic in Europe and the steppe as far as Altai; extending their occupation for extended periods to southwest Asia. Denisovans are thought to be endemic to steppe and Sunda. A hominin population descended from H. erectus was endemic in Asia and Sunda. Further out, H. floresiensis (“the Hobbit”) survived in insular isolation on Flores. Reich reports that there was at least one other “ghost population” in Eurasia whose fossils have yet to be identified but whose genetic signal is evident in ancient DNA from admixture. H. sapiens was endemic in Africa where it shared the continent with yet other ghost populations. (On population structure in prehistoric Africa see Scerri et al. 2018).

So what is the evidence for modern behavior for these taxa during the Upper Pleistocene? So far we only have reliable archaeological evidence for Neanderthals and anatomically modern humans.

Gamble (2013) characterizes the Neanderthal dispersal in western Eurasia as ‘an adaptive radiation based on projectile technology and enhanced carnivory with prime-age prey the target’ c. 300-200 Ka. Not only did Neanderthals invent hafting, Upper Pleistocene assemblages associated with Neanderthals exhibit disproportionately higher frequencies of prime age prey compared to anatomically modern humans, suggesting that they were more competent hunters. In the same assemblages we find the first evidence of managed fire. In a separate analysis, Zilhão (2007) notes that

Chemical analysis of two fragments of birch bark pitch used for the hafting of stone knives and directly dated to >44 ka 14C BP showed that the pitch had been produced through a several-hour-long smoldering process requiring a strict manufacture protocol, i.e., under exclusion of oxygen and at tightly controlled temperatures (between 340 and 400ºC) (Koller et al., 2001). The Königsaue pitch is the first artificial raw material in the history of humankind, and this unique example of Pleistocene high-tech clearly could not have been developed, transmitted, and maintained in the absence of abstract thinking and language as we know them; it certainly requires the enhanced working memory whose acquisition, according to Coolidge and Wynn (2005), is the hallmark of modern cognition.

Gowlett (2010) suggest that Neanderthal hearths were a game changer. Not only did fire act as an external stomach by breaking down enzymes in meat thus reducing gut loads, it also extended the day for social interactions. More compelling evidence on Neanderthal social life comes from ‘super sites’ recognized from assemblages with massive deposits of debris accumulated over very long periods of time. ‘They marked,’ Gamble (2013) notes,

… a shift in hominin imagination from conceiving the world in a horizontal manner to stacking it vertically – a representation of accumulated time. … Once established, the super-site niche set the stage for thinking differently about places. They now formed part of the hominins’ distributed cognition.

The supersites suggest that Neanderthals had indeed ‘discovered society’; surely an important consideration in the Neanderthal question—whether they were biologically capable of modern behavior. Surely if they had complex social lives it becomes hard to classify them as subhuman.

However, the litmus test for artifact assemblages is evidence of symbolic storage. Shea (2011) seconds João Zilhão and Francesco d’Errico that ‘finds of mineral pigments, perforated beads, burials and artifact-style variation associated with Neanderthals challenge the hypothesis that symbol use, or anything else for that matter, was responsible for a quality of behavioral modernity unique to Homo sapiens.’ Caspari et al. (2017) note that,

Neanderthals were capable of complex behaviors reflecting symbolic thought, including the use of pigments, jewelry, and feathers and raptor claws as ornaments. All of this is probably evidence of symbolic social signaling, complementing the evidence of complexity of thought demonstrated by the multi-stage production of Levallois tools. Neanderthals were human.

But problems for the recent mutation hypothesis don’t end with behaviorally modern Neanderthals. Shea (2011) explains how any joint examination of the diachronic and synchronic pattern of the actual behavior of anatomically modern humans as they dispersed from Africa undermines the recency hypothesis:

Evidence is clear that early [anatomically modern] humans dispersed out of Africa to southern Asia before 40,000 years ago. Similar modern-looking human fossils found in the Skhul and Qafzeh caves in Israel date to 80,000 to 120,000 years ago. Homo sapiens fossils dating to 100,000 years ago have been recovered from Zhiren Cave in China. In Australia, evidence for a human presence dates to at least 42,000 years ago. Nothing like a human revolution precedes Homo sapiens’ first appearances in any of these regions. And all these Homo sapiens fossils were found with either Lower or Middle Paleolithic stone tool industries.

Shea (2011) appreciates the evidence from Howiesons Poort c. 90-70 Ka and other sites of early ‘efflorescences’. But that deepens the paradox. For

… [if] behavioral modernity were both a derived condition and a landmark development in the course of human history, one would hardly expect it to disappear for prolonged periods in our species’ evolutionary history.

Habgood (2007) examined the evidence for modern behavior in Pleistocene Sahul and concludes:

The proposal by McBrearty and Brooks (2000) and Mellars (2006) that the complete package of modern human behaviour was exported from Africa to other regions of the Old World and ultimately into Greater Australia between 40–60ka BP is not supported by the late Pleistocene archaeological record from Sahul….

O’Connell and Allen (2007) put things more bluntly. Sahul’s archaeological record, in their opinion, ‘does not appear to be the product of modern human behaviour as such products are conventionally defined.’ Equally sensitive observations may be made for southern Africa, southern Eurasia, eastern Eurasia, or Sunda, where modernity is not evident until 30-10 Ka, ie tens of thousands of years after the arrival of “moderns” (if not later).

The greatest problem with the recent mutation hypothesis might very well be that it necessarily implies a modern-nonmodern dichotomy since you either have the trait for generative grammar or you don’t. In that sense it is very much like Krugman’s It theory of global polarization. This dichotomy pronounces not only other archaic hominin but also many anatomically modern humans to be nonmodern and hence not fully human well into the ethnographic present. Indeed, it has proven exceedingly difficult to quarantine Neanderthals from anatomically modern humans for there is no hyperplane in the space of behavioral traits that includes both hunter-gatherer societies in the ethnographic present and excludes Neanderthals. As Zilhão (2011) notes,

Many have attempted to define a specifically “modern human behavior” as opposed to a specifically “Neandertal behavior,” and all have met with a similar result: No such definition exists that does not end up defining some modern humans as behaviorally Neandertal and some Neandertal groups as behaviorally modern. [Emphasis added.]

In a comment on Henshilwood and Marean (2003), Zilhão captures the essence of the conundrum induced by the implied dichotomy:

… the real problem is that (1) archeologically visible behavioral criteria designed to include under the umbrella of “modernity” all human societies of the historical and ethnographic present are also shared by some societies of anatomically nonmodern people and (2) archeologically visible behavioral criteria designed to exclude from “modernity” all known societies of anatomically nonmodern people also exclude some societies of the historical and ethnographic present.

We can think of this as “merely” a political issue and tell ourselves that we should “just do the science; let the chips fall where they may.” Or we may ask ourselves if such essentialist schema are doing as much explanatory work as they are supposed to.

At this point we can anticipate the following rebuttal:

Why is it such a big deal if the efflorescences get extinguished or paleo-demes fail to exhibit modern behavior for tens of thousands of years after the posited mutation? Since the capacity for generative grammar is uniform across the species, no denial of modernity is implied since any evidence of modern behavior by one paleo-deme in the species is enough to confirm the potentiality for modern behavior for all demes in the species. These may appear in a staggered manner. But that can be easily explained by adding that the capacity for generative language may only be a necessary condition for modern behavior and that other things would also have to fall into place before any ‘full flowering’ obtains.

This is a fair response. And we speculate that something like this was indeed the case. But notice that in light of the evidence relayed above, the “one-strike-in” rule for taxa immediately implies that Neanderthals were behaviorally modern as well. This is indeed what we have been arguing all along. And if that is the case then the mutation had to have occurred way, way earlier—at least hundreds of thousands of years before the 80 Ka posited by Berwick and Chomsky (2016). This interpretation is in line with the conclusions offered in a recent volume edited by Brantingham et al. (2004) published under the title The Early Upper Paleolithic beyond Western Europe:

If there is a common evolutionary cause, phylogentic [sic] or otherwise, it is rooted much deeper in evolutionary time and is largely independent of the events tracked in the Middle-Upper Paleolithic transition. [Emphasis added.]

Zilhão (2007) concurs with the last assessment offering 400 Ka as a plausible date of arrival of the biological wetware for modern behavior. But if that is true, then it raises the daunting question of why said mutation did not generate an adaptive radiation. Perhaps it corresponds to the adaptive radiation associated with the dispersal of H. heidelbergensis 600-400 Ka (‘Out of Africa II’) that Gamble (2013) ties to mastery of controlled fire. But that possibility is unlikely to satisfy those who seek to explain ‘the revolution that wasn’t’ (McBrearty and Brooks 2000; the reference is to “the Human Revolution”) for a genetic mutation c. ~400 Ka cannot then explain the adaptive radiation evident c. ~50 Ka.

Bar-Yosef (2002) suggests that a better analogy for modern behavior would be the Neolithic Revolution, with its staggered appearance. Indeed, what is manifest is that global polarization made its first appearance not in the Neolithic but during the Upper Pleistocene. The diachronic and synchronic patterns revealed by archaeology are consistent with the framing of modern behavior during the Pleistocene as an emergent form of sociocultural complexity that makes a staggered appearance and therefore necessarily generates global polarization.  It was the First Great Divergence whose roots may be sought perhaps in biogeography.

The weight of biogeography is well captured by Bailey (1960)’s Effective Temperature (ET) that measures both the basic thermal parameter of the ecoregion and the length of the growing season. Binford (2001) shows how ET structures the lifeworld of hunter-gatherers; opening some doors while closing others; rigging the dice in favor of some and against others. For as Gamble (2013) notes mercilessly, ‘the low latitudes were good areas to disperse away from …’. Proctor (2003) acidly calls this discourse “Out of Africa! Thank God!” But is there any doubt that escaping the tyranny of the isotherms was an unambiguous advantage? as it has been ever since? Our position is that the antiracist suspicion of biogeography is completely mistaken. Biogeography is not racist but rather an alternative to racialism as an explanatory schema for global polarization.

What of the fate of the Neanderthals? The weight of the evidence suggests that they were not wiped out but rather absorbed into the much larger incoming populations of anatomically modern humans. They may have been as dynamic as H. sapiens. But there was a decisive difference. Our race had much longer life histories. This suggests that the adaptive radiation witnessed c. ~50 Ka may be related to runaway K-selection in anatomically modern humans. With grandparents around to pass on know-how and know-where the fidelity of intergenerational transmission of acquired innovations must have been higher. And larger populations could be expected to generate more innovations. It may be demographic and life history variables that explain the diachronic and synchronic patterns of our deep history.