1. The Symbolic Species

The list of unique human features is almost endless, mostly involving physical alterations that stem directly or indirectly from our unusual form of upright locomotion. But the attribute that makes us feel so different from all the other living organisms around us is an intangible one, involving the way we process information in our brains. Alone among all creatures on the Earth today, we human beings are symbolic, meaning that we mentally disassemble the world around us into a vocabulary of abstract symbols - which we can then recombine in our minds to create alternative versions of reality. We cannot entirely escape reality, as any surviving victim of a natural disaster will tell you; but unlike other organisms which respond, with greater or lesser sophistication, to stimuli impinging upon them, we human beings are able to live for much of the time in the world as we re-create it in our heads. The fact that each one of us will re-create it slightly - or even radically - differently is, of course, at the root of the unusual complexities of human society, and of the need every society exhibits to have elaborate rules and procedures governing the behaviors of its members.

Today, then, we are most strikingly symbolic (see, for example, Lock and Peters, 1996). But the way in which we otherwise fit so seamlessly into the great Tree of Life on Earth indicates unequivocally that we are descended from a non-symbolic ancestor. Unique features are famously hard to explain; and indeed, the only reason we have for thinking that any earthbound form could ever have become symbolic, is that we so obviously did. But if we are ever to form a proper understanding of ourselves, and of just how we fit into Nature, we will have to understand exactly how it was that we arrived at our entirely unprecedented cognitive state.

This is complicated because not only cognitive states, but also the behaviors they generate, do not leave direct tangible traces. All we have by which to trace the history of our precursors' transformation from a more or less run-of-the-mill primate to the totally unmatched physical and cognitive entity that Homo sapiens is today, is our fossil and archaeological records. The first of these consists of the mineralized bones and other direct traces of themselves left by earlier members of our zoological family Hominidae (or subfamily Homininae; it makes no practical difference); the second consists principally of the artifacts these hominids made, and the ways in which they were strewn both around the landscape and the specific spots at which they were found. And in each case, the indicators to hand must stand as indirect proxies for behaviors and cognitive states that must themselves necessarily be inferred. Such inferences are often arguable, and in many cases they are not directly testable. Still, the overall pattern of change over most of hominid evolution is pretty clear, and is briefly outlined below, following an attempt to clarify an initial cognitive state.

2. The Starting Point

One of the problems that paleoanthropologists face that paleontologists of other stripes do not, is that Homo sapiens is the lone representative of its family in the world today. We no longer have anything close to compare ourselves with. But while our closest living relatives, the chimpanzees and bonobos, are actually fairly remotely related to us, they nonetheless bear sufficient cranial and brain-size resemblances to our earliest hominid ancestors that we can reasonably view them as approximate cognitive models for those ancestors.

In this context, we can concede right away that cognitively the apes are very impressive. Indeed, barely a week seems to pass without the discovery that one or another of them displays yet another behavior that we had thought unique to ourselves. An impressive recent example is the use by savanna chimpanzees of wooden "spears" to impale sleeping bushbabies (Pruetz and Bertolani, 2007). What's more, "ape language" experiments have also shown that these primates are capable of recognizing and responding not only to individual visual and auditory symbols, but to limited combinations of them (Cohen, 2010). Still, there can be no doubt that apes are nonsymbolic in the human sense, and the cognitive scientist Daniel Povinelli suggests one limiting factor is that "chimpanzees rely strictly upon observable features of others to forge their social concepts … they do not realize that there is more to others than their movements, facial expressions, and habits of behavior" (Povinelli, 2004: 33). From these and other observations, Povinelli concludes that very early hominids did not "reason about unobservable things" and lacked any "notion of causation" (Povinelli, 2004: 34). And while this negative deduction involves a long chain of inference, it nonetheless seems to be as robust a statement as we can currently achieve about the cognitive condition of our earliest ancestors.

3. Patterns of Change in Human Evolution

The very earliest hominid fossils known are in the seven to four million year (myr) age range. Mostly fragmentary, they make a quite oddly assorted group, and perhaps their most important function is to demonstrate that from the very beginning the hominid family tree has been bushy much like that of any other successful mammalian family. It is highly unusual for Homo sapiens to be the only hominid on Earth (Fig. 1). Among other things, this means that we need to resist the temptation to reconstruct the history of our species by extrapolating it back into the past to reveal a sort of singleminded slog from primitiveness to perfection, a process of fine-tuning over the eons. Earlier hominids were emphatically not simply junior-league versions of us.

The earliest well-documented hominids are classified as Australopithecus afarensis. Fossils of this species are known from several eastern African sites in the 3.6 to 3.0 myr time range, during which the African climate was drying and dense forests were giving way to woodland. When on the ground these small-bodied hominids walked bipedally, presumably because they were already most comfortable holding their trunks upright while climbing arboreally. Still, they were broad-hipped and short-legged, and many features of their upper bodies indicate retained agility in the trees. Similarly, their brain sizes and cranial proportions remained comparable to those of chimpanzees, hence their frequent description as "bipedal apes." Still, ecologically there was a significant difference. For while chimpanzees living in open environments today still eat basically the same fruit-based diet as their forest-dwelling fellows, it is clear that the early hominids had expanded their dietary repertoire to include the new resources offered by woodland and savanna environments. These probably included not only tough tubers and rhizomes, but probably also small mammals such as hyraxes, as indicated by analysis of stable carbon isotopes preserved in their teeth (Sponheimer and Lee-Thorp, 2007).

The pursuit of such resources along the forest-edges and in the expanding woodlands exposed the "bipedal apes" to significant hazards of predation; and the potency of this new factor suggests that in reconstructing their social organization we should turn not to today's forest-dwelling apes, but to other higher primates that live in similarly dangerous habitats, namely, the baboons and macaques. These highly social primates live not in small groups as apes do, but in extremely large ones that have complex social hierarchies, and are spatially organized to protect the crucially important but vulnerable core of females and infants (Hart and Sussman, 2009). These groups are, however, sufficiently flexible to break up into subunits for daily foraging out in the open as necessary, while coming together in trees or on rock-faces for nocturnal protection. Physically the bipedal apes were supremely well-equipped for a lifestyle of this kind, that spanning grassland to forest - and we know that theirs was not a "transitional" adaptation, but a stable and successful one that remained essentially unaltered for millions of years, even as new species came and went.

We have no reason to suspect that the earliest A. afarensis differed significantly in their underlying cognitive processes from what we see among the social great apes today - at least until, at about 2.5 myr ago, we find evidence of the first stone tools - simple sharp flakes knocked off rock "cores." Here is evidence of a huge cognitive leap that not only involved appreciation of the qualities of materials, but that also demanded planning and foresight: appropriate rocks were carried long distances before being made into tools (Klein, 2009). Although some claimed "early Homo" fossils are known at 2.5 myr, the earliest stone tools are almost certainly the work of Australopithecus - and there are even intimations - in the form of butchery marks on animal bones - that such tools were being made substantially earlier (McPherron et al., 2010). Thus, right from the beginning, a consistent theme is established: the fossil and archaeological records are out of phase. Still, the new ability to cut and dismember carcasses must have hugely affected the lives of the first toolmakers, and it may signal a significant acceleration of the tendency to include animal fats and proteins in the diet. Sadly, the evidence doesn't allow us to say anything about how this technological advance might have reflected change relative to the apes in the toolmakers' subjective experience of the world.

The first unequivocal members of our genus Homo, the earliest hominids with basically modern body size and structure, are known in East Africa at around 1-9 to 1.6 myr ago. They had reduced faces in concert with brains that were larger than those of bipedal apes, although they were not much more than half the size of our brains today. And while their species, Homo ergaster, is initially found in association with stone tools of archaic type, these hominids must have undergone a huge lifestyle change.

This is because they were physically committed to the open savanna, where the range of dietary resources available, and the ambient conditions, were radically different from anything experienced by their forebears. Indeed, it has been argued not only that an increase in animal protein intake was mandated at this point by the energy-hungry larger brain, but that meat - and plant foods, as well - must have been cooked to increase the availability of the nutrients contained (Wrangham, 2010). It is even possible that the reduction in metabolically-expensive gut tissue typical of the genus Homo was an inevitable price of increasing brain size. However this may be, domesticated fire would certainly have been valuable in discouraging a daunting carnivore fauna, especially at night; and it would also have provided a physical focus for the group, something that would in turn have had huge social ramifications. All this makes an attractive argument, but for the time being it remains a circumstantial one, since there is no tangible evidence for the domestication of fire before about 800 thousand years (kyr) ago (Goren-Inbar et al., 2004). All we can say with any confidence at present is that with the new body form we are seeing hominids transitioning from prey species to a (still vulnerable) role as secondary predators. Like us the new hominids were slow; but they had enormous endurance under the hot sun, which may have enabled them to obtain fleeter by wearing them down. And, as predators, they would necessarily have been thin on the ground, suggesting that earlier large group sizes had become drastically reduced and perhaps intensifying social bonds among group members.

A new kind of stone tool appears at about 1.5 myr ago. "Handaxes" were large implements deliberately formed to a standard symmetrical teardrop shape, implying for the first time the existence of a "mental template" in the minds of the toolmakers (Schick and Toth, 1993). Clearly, another cognitive advance is implied; but alas this innovation tells us little about Homo ergaster's wider experience of the world. Still, and probably very significantly, we have nothing at all at this stage to suggest any kind of symbolic behavior. Indeed, there is nothing in any aspect of Old Stone Age technology, right up until the very last Ice Age, into which we can confidently read symbolic input. Craftsmanship, yes; craftiness, yes; even an intuitive aesthetic appreciation. But not symbolism.

This continued to be true with the advent of Homo heidelbergensis, the first cosmopolitan hominid species. Appearing in Europe and Africa at about 600 kyr ago, this hominid spread as far east as China, and its tenure witnessed the first building of artificial shelters, the regular domestication of fire, the first hafting of stone tools into compound implements, and the first fabrication of javelin-like throwing spears (see overview in Tattersall, 2009a), all documented at about 400 kyr ago. Somewhat later, in the 300-200 kyr span, came the next radical invention in stone tool-making. This was the "prepared-core" tool, produced by fashioning a stone core on both sides until a single blow could knock off a flake of predetermined form. Brains of the robustly-built Homo heidelbergensis were within the modern size-range, though a little under the modern average; but again, this species bequeathed us nothing that we can unequivocally identify as the product of a symbolic mind. Homo heidelbergensis was clearly an accomplished hunter and exploiter of its environment; but the most we can currently say of it is that it was doing what its predecessors had done, but a bit better. No doubt these hominids had a very sophisticated intuitive intelligence, and complex ways of communicating involving vocalization, gesture, and body language; but we have no grounds for believing it communicated or mentally processed information in the ways that we do - even in rudimentary form. It is thus particularly frustrating that we have such extreme difficulty in seeing through the eyes of any form that does or did not share our particular human mindset. Homo heidelbergensis is often reckoned to be the progenitor of the lineages that lead to Homo neanderthalensis on one hand (in Europe), and to Homo sapiens on the other (in Africa). Whether or not this is the case, in the later part of the Pleistocene epoch we can identify several independent hominid lineages in which brain size was independently increasing. It was happening in Europe, in the Neanderthal lineage; in Africa, in the lineage that eventuated in Homo sapiens; and in eastern Asia, in the Homo erectus lineage. Clearly, there was something about the genus Homo that predisposed its members to metabolically-expensive brain expansion, regardless of geography or environment; and understanding just what this something was, will be crucial to comprehending how we human beings became the extraordinary creatures we are. But it was clearly not linked to the specific and unusual way in which we process information in our brains, because for neither the Neanderthal nor the Homo erectus lineage is there any firm evidence that symbolic cognition was ever achieved.

The Neanderthals are particularly fascinating because they are the best-known of all extinct hominids, and thus provide the most reliable yardstick we have by which to measure our own uniqueness. Homo neanderthalensis occupied Europe and a wide swath of western Asia between about 200 kyr and 30 kyr ago, and had a brain as big as ours. Physically distinctive (from us; from the neck down it was probably a fairly standard later Homo species: Tattersall, 2009b), this hominid was clearly a skilled hunter of large animals and exploiter of its (often severe) environments (see discussion and references in Tattersall, 2009a). Neanderthals were an accomplished practitioners of the prepared-core tool-making method, and they buried their dead (at least occasionally, and without fanfare) while at the same time practicing a chillingly prosaic form of cannibalism (Carbonell et al., 2010). Clearly, the Neanderthals communicated subtly and effectively with each other, and had a keen understanding of the difficult world around them. But, again, over the vast expanse of time and space they inhabited these hominids left no unequivocal proxy evidence of symbolic thought processes; and it is probably telling that they were entirely eliminated, in a few short thousand years, by the fully symbolic Homo sapiens who began trickling into their territory at about 40 kyr ago. These were the Cro-Magnons, the creators of such stupendous cave art sites as Chauvet, Lascaux and Altamira: beings whose lives were totally drenched in symbol.

4. Becoming Human

Modern Homo sapiens fossils begin appearing in Africa (without clear fossil antecedents) at about 200-160 kyr ago. Archaeological associations are scanty and modest (Klein, 2009). Better archaeological contexts (Klein, 2009) exist for the first anatomically modern humans outside Africa. These are found at around 100 kyr ago in Israel which also, at least intermittently, hosted Neanderthals between about >130 kyr and 45 kyr. And throughout this period, the stone tool assemblages associated with the two kinds of hominid are virtually identical. There is no reason whatever to suspect any cognitive difference between the two in this period. Neither was symbolic. This actually agrees with the record in Africa, because the first intimations of symbolic behavior in that continent are found subsequent to about 100 kyr ago, in the form of geometrically-engraved ochre plaques and complex multi-stage technologies reported from sites on the South African coast in the period around 75 kyr ago (Henshilwood et al., 2004; Brown et al., 2009). Even earlier intimations exist in the form of pierced shell "beads" from sites both in the north and the south of the continent; but such expressions are more arguably symbolic, and are still in the <100 kyr range. Thus, modern human anatomy appears to have been achieved significantly before modern symbolic behavior patterns.

So what happened? Most plausibly, after something like 400 myr of vertebrate brain evolution, the neural substrate permitting symbolic thought had been produced, in an already near-enabled brain, as a byproduct of the apparently radical developmental reorganization that gave rise to Homo sapiens as an anatomically distinctive entity (Tattersall, 2009b). Whatever its exact nature, this innovation lay fallow until its new use was "discovered" by its possessors, through the effect of some necessarily cultural stimulus. However radical its result, the evolutionary mechanism of exaptation that must have been involved here is not in the least unusual; after all, the ancestors of birds had feathers for million of years before using them to fly. So what was that cultural stimulus? The most plausible candidate is the invention of language, which is almost synonymous with what we experience as thought today. Like thought, language involves creating symbols in the mind, and rearranging them according to rules to generate an infinite number of possible statements from a finite vocabulary. What's more, it is known that structured language can be readily and spontaneously invented (Kegl et al., 1999); and, as an externalized attribute, language is highly likely to spread rapidly within populations already biologically predisposed for it. Once symbolic thought had been "kicked" into existence in this way, newly symbolic Homo sapiens was able to envision alternative worlds, and thus to plan in an unprecedentedly complex manner. It acquired that sense of "self" that Povinelli found so conspicuously lacking in chimpanzees. This opened up a huge cognitive gulf between our species and even its closest contemporaneous relatives (the Neanderthals, Homo erectus, its own ancestor). Complex as they undoubtedly were, these "old-style" hominids were cognitively limited to responding, in however sophisticated a manner, to the outside world as it was presented to them by Nature. They were unable to remake the world in their minds, as Homo sapiens now could. At some point after acquiring symbolic thought in its parent continent Homo sapiens left Africa, and its rapid spread through the Old World, and eventually to the New has been exhaustively recorded by molecular biologists (see overview by DeSalle and Tattersall, 2008). Its nearest (and apparently not so dearest) residing in those territories never stood a chance.