For a highly egocentric species that is exquisitely conscious of both its unusual anatomical structure and its unique way of mentally processing information, the otherwise self-obsessed Homo sapiens has shown remarkably little interest in formally defining itself relative to the rest of the living world. Back in the eighteenth century the Swedish savant Carolus Linnaeus, father of the system of classifying animals and plants we still use today, established the practice of using particular biological features to define each genus and species in his classification. Yet, in the case of our genus Homo—and only in that case—Linnaeus departed from this sensible procedure, casually advising his readers, nosce te ipsum (know thyself).

Even at the time, this exhortation toward introspection might hardly have appeared the ideal approach to biological self-definition. But in practice few cared: after all, Linnaeus’s lexicographer contemporary Samuel Johnson was satisfied with defining man as a human being, and human as having the qualities of a man. Quite simply, everyone then thought they knew that human beings were so different as to require no definition at all.

Of course, in the eighteenth century, nobody had any idea that creatures had once lived that were structurally intermediate between us modern humans and the great apes, our (then barely known) closest living relatives. People lived in a world in which their own distinctiveness was self-evident. But this blissful ignorance could not continue forever, and when it was finally dispelled in the mid-nineteenth century by the emergence of evolutionary thinking, the entire perspective on humanity’s place in nature changed.

You might thus be tempted to imagine that, in the century and a half since Charles Darwin pointed out that we are joined to the rest of nature by common ancestry, science might have begun to make some progress toward a biological definition of the human genus. But if so, you would be doomed to disappointment. Scientists are still arguing vehemently over which ancient fossil human relatives should be included in the genus Homo. And they are doing so in the absence of any coherent idea of what the genus that includes our species Homo sapiens might reasonably be presumed to contain.

Understanding why this should be so requires a brief return to the basics of Linnaeus’s classificatory system. That system is an inclusive one, consisting of a hierarchy of groups-within-groups. The most fundamental unit in the hierarchy is the species, usually defined by reproductive exclusivity. Basically, species are the largest populations of organisms within which breeding can freely take place. Closely related species—those sharing the same common ancestor—are grouped into the same genus; genera are grouped in turn into families, families into orders, and on up the line. In principle, then, a genus is a pretty straightforward thing: it is a grouping of species that all descend from the same common ancestor.

But in practice there is a problem: the splitting of lineages that produces those groups of species also produces a branching system of relationships. In other words, the descendants of the ancestral species themselves divide, to produce ever-larger groupings of descendants that are increasingly remotely related. As a result, not all species in any genus will be equally closely related by descent. This means that how inclusive you want your genus to be is an entirely arbitrary matter about which there may be legitimate disagreement. Species have a reasonably objective biological reality that is grounded in the dynamic that exists among their members. Genera, on the other hand, are purely historical constructs.

Fortunately, in the living world genera tend to be fairly readily recognizable empirically, and there is actually a lot less argument about the boundaries of living genera than there is about those of living species. But whereas there is a great deal of information about contemporary organisms upon which classifiers can draw, in the paleontological realm the information available is typically much more limited. The fossil record, mainly consisting of mineralized bones and teeth, is a very pale reflection of the once-living world it represents, and it is rare indeed to find an articulated skeleton that will reliably tell you which elements of the individual go together. Worse yet, the fossil bones themselves are more often than not broken and incomplete. As a result, attributing fossils to a species or to a genus can often be an extremely tricky matter involving a lot of subjective judgment. This fairly harsh reality provides much of the background against which attempts to recognize fossil members of the genus Homo need to be understood. But there is also a dimension of history and of perspective that is particularly significant in the context of paleoanthropology, the study of human evolution.

For many years, the analysis of the human fossil record was largely the province of human anatomists. By training, these scholars were exquisitely attuned to small variations on the basic Homo sapiens theme, but at the same time they were also ignorant of the larger patterns of diversity in nature with which other kinds of paleontologist routinely dealt. Because of this, they tended to use zoological names not as a means of expressing and organizing that diversity, but simply as designators of individual fossil specimens. As a result, by the mid-twentieth century there were as many as a dozen different genus names in common use for a relatively modest collection of known human fossils, most of them Neanderthals.

Ultimately, this was an untenable situation, and in 1950 the influential ornithologist Ernst Mayr, who it must be pointed out had probably never seen an original human fossil in his life, took it upon himself to put paleoanthropologists firmly in their place.1 Mayr was one of the architects of the new evolutionary synthesis, which over the preceding decades had married genetics, systematics, and paleontology to provide the first integrated account of the evolutionary process. In essence, the synthesis saw evolution as a gradual thing. Over time, one species gave rise to the next, as genetic novelties accumulated under the guidance of natural selection (basically, differential reproduction, as better adapted individuals out-reproduced inferior ones). And accordingly, Mayr declared that not only had there been but one single genus—Homo—in the entire history of the hominid family, but that it had contained a succession of only three species.

The earliest of these was his Homo transvaalensis. This was broadly equivalent to what we know as australopiths today: small-bodied, archaically-proportioned bipeds with ape-sized brains and large faces. These hominids, which flourished in Africa in the period between about 4 and 1.5 million years ago, are not infrequently described as bipedal apes. Homo transvaalensis then gradually evolved into Homo erectus, with more modern body proportions and a larger brain, and this yielded in turn to Homo sapiens, which Mayr saw as embracing the large-brained but structurally primitive Neanderthals as well as ourselves.

Lumping this huge variety of hominids into the single genus Homo was an act of systematic extremism that in practice made it impossible to characterize the human genus in either anatomical or behavioral terms. Yet the paleoanthropologists capitulated instantly. Operationally, the units of paleoanthropological analysis rapidly became individual fossils, or informal groupings such as the Neanderthals. Zoological names, genera included, were largely eschewed.

The first major complication arose in 1964, when Louis Leakey and his colleagues described the new species Homo habilis from Tanzania’s Olduvai Gorge.2 The fossils concerned more closely resembled australopiths from South Africa than anything else then known, but they were associated with primitive stone tools, and Leakey was very much a devotee of the then-fashionable notion of “Man the Toolmaker.” Leakey also—accurately—claimed that his fossil had a brain slightly bigger than typical for australopiths, but the key attribute of the genus Homo was nonetheless a behavior: stone tool making. Anatomy and biology thus once again faded into the background.

It is worth noting that subsequent research has pretty definitively associated the earliest known stone tools (some 2.5 million years old, now possibly more) with australopiths.3 What is more, over the following decades the name Homo habilis become a receptacle for a rather odd grab bag of African hominid fossils, most of which have a pretty archaic look to them. The earliest Homo habilis fossils from Olduvai have been dated to 1.8 million years ago, which happens also to be the age of the amazing assemblage of hominid remains found from 1992 on at the site of Dmanisi in the Republic of Georgia.4 These fossils, also associated with crude stone tools, have been variously attributed to Homo habilis, to Homo erectus, and to the new species Homo georgicus, and while they show an amazing range of size and morphology, what they all have in common is remarkably small brains, hardly larger than those of australopiths. This is a conspicuous contrast with Homo sapiens, the standard-bearer of the Homo genus—and the yardstick by which all other putative members have to be judged—which has a brain of more than twice the volume. Yet, while they have vacillated as to the species identity of their fossils, the Dmanisi researchers have never wavered in attributing them to the genus Homo—even though doing this made it yet more difficult than it had been to arrive at any coherent morphological or behavioral characterization of the genus.

Not until 1999 did anyone express significant concern about the wholesale cramming of miscellaneous hominid fossils into Homo that had deprived the genus of any anatomical rationale. In that year, the paleoanthropologists Bernard Wood and Mark Collard argued that Homo should be redefined to include only fossil species that met certain criteria of structure and relationship.5 Under these criteria, Wood and Collard saw only forms demonstrably more closely related to Homo sapiens than to the australopiths as admissible into Homo. What’s more, they required that all putative Homo should show the prolonged developmental timetable that finds its apogee in modern humans, resemble Homo sapiens more closely than australopiths in skull proportions, and possess body sizes and structures compatible with being committed terrestrial bipeds like us. The diminutive australopiths, in contrast, still maintained a lot of ancestral tree-climbing in their locomotor repertoire.

How wise it was to cram the entire huge variety of fossil hominids known in 1999 into a simple australopith-vs-Homo dichotomy remains debatable. But there is no doubt that Wood and Collard’s suggestions marked a huge advance toward arriving at a reasonable biological definition of the genus to which we Homo sapiens belong. Since no fossils more than two million years old are currently known that fit the Wood/Collard criteria for Homo membership, all members of the resulting morphologically-coherent genus Homo appear as the fruit of a common ancestor that lived in Africa a little under two million years ago. Its descendants include the species Homo ergaster (the earliest African form), Homo erectus (indigenous to eastern Asia), the cosmopolitan Homo heidelbergensis, the European Homo neanderthalensis, and the African-derived Homo sapiens. All of these have at least modestly large brains, and basically modern body proportions. Leakey’s Homo habilis doesn’t make the cut, and neither does a form of similar age from Kenya that has been called Homo rudolfensis. As to morphology, the ancestor itself (possibly Homo ergaster, but more likely as yet undiscovered) necessarily already possessed the distinctive suite of anatomical features in both the skull and the body that is shared to varying extents by its descendants. Members of all Homo species as thus defined would have been readily recognizable on the landscape as hominids of the same general kind, despite differences in brain size and cognitive style.

These ancestral adaptations clearly reflected a radical shift, away from an ancient way of life that included feeding and sheltering in the trees, and toward one based on exploiting the resources offered by more open environments. Such habitats were actively expanding over the life of the hominid family as African climates became drier and more seasonal. One major new resource these environments offered was the carcasses of mammals and other vertebrates, and it was these that most probably provided the high-energy diet necessary to support the dramatic expansion of the energy-hungry brain so abundantly documented in the fossil record of Homo subsequent to about two million years ago. Along with increasing prowess in hunting, and the butchering presumably associated with it, brain enlargement seems to have been the major theme in the evolution of our genus, occurring independently in several lineages within it—and not just in our own.

Sadly, Wood and Collard’s sage advice has been widely ignored over the almost two decades since they wrote, in favor of the continued search for the earliest Homo that Leakey had begun at Olduvai. Paleoanthropology may have been a little unfairly described as discovery-driven;6 but, once Leakey’s allocation of the Olduvai fossils to Homo had become generally accepted, the race was clearly on to discover ever more ancient Homo fossils, usually via some form of demonstration that the hominid in question was not an australopith. Since the 1970s, a succession of mainly fragmentary fossils in the 2.0 to 2.5 million-year range has been attributed to early Homo, the inclusion of each one broadening the morphological range that appeared acceptable within Homo, and thereby making it easier to include others—even as the genus itself became progressively harder to define in any terms whatever.

The latest entrant in the early Homo stakes is a 2.8 million year-old mandible fragment from a place in Ethiopia called Ledi-Geraru.7 A number of its characteristics were put forward to support the allocation of this fossil to Homo. The genus, it was then consequently argued, had very deep roots in time, and showed great diversity early on. Again, however, the features concerned were largely ones that distinguished the mandible from those of contemporaneous australopiths, rather than attributes that bound it to any coherent morphological or genealogical concept of the genus Homo. This was inevitable, of course, since at 2.8 million years ago the morphologically coherent genus Homo envisioned by Wood and Collard evidently had yet to evolve.

The problem lay in the attempt to divide early hominids into Homo vs australopiths, rather than admitting that the morphological diversity actually seen way back into the early hominid fossil record would best be organized by recognizing more genera.8 If this were done, it might be more readily seen that it is the larger group which contains Homo that has deep roots in time, rather than that genus itself, which encapsulates a grouping of closely (but not equally) related species whose common ancestor existed significantly later than Ledi-Geraru times.

As we have already seen, via the splitting of lineages nature produces a branching genealogy of species that multiplies through time (though most sooner or later become extinct). Reconstructing the complex history of branching in such genealogies is poorly served by excessive lumping together of their components, which themselves become increasingly difficult to define in the process.

All of this points in one direction: to the urgent need for a compact, phylogeny-based, consensus-based definition of the genus Homo of the kind already sketched by Wood and Collard. If a generally-agreed morphological definition had been available in 2004, when the famous hobbit from the Indonesian island of Flores was reported, the scientists who described it might not have felt obliged to name that tiny-brained and oddly-proportioned hominid Homo floresiensis, thereby hugely complicating the definition of Homo. The Flores hominid is patently not an australopith.9

The latest example of how unhelpful this kind of reasoning can be is the new South African hominid recently described, to enormous press fanfare, as Homo naledi.10 This is one of the most extraordinary paleoanthropological discoveries of recent years, the fossils concerned consisting of a mass of broken hominid bones lying on the floor of an almost inaccessible limestone cave. So far some 1550 specimens of naledi have been recovered from the Dinaledi Chamber, as the findspot is called, with many more reportedly to come. The fragments analyzed so far represent the remains of at least 15 different individuals, giving us a rare peek at the demographics of an archaic hominid species. Their geological age has yet to be determined, their circumstances of deposition having so far resisted available techniques of dating. This is a little frustrating, of course; but at least it does give us a chance to weigh the anatomy of the fossils dispassionately, without being influenced by their age.

The group of scientists who found this extraordinary trove has acted with laudable (and very unusual) speed in presenting the fossils to the public and to their colleagues; they were first spotted in late 2013, and by September 2015 preliminary descriptions and 3D images were already available. This inevitably means that the Dinaledi scientists’ conclusions about this material are necessarily provisional; but it is already clear that naledi possesses an unusual combination of morphological attributes. Males of the species are said to stand about five feet tall, females a little less. This puts stature just within the lower boundaries of Homo sapiens, and at the very high end for an australopith. The pelvis is flared like that of an australopith, while the feet seem to have been remarkably modern in comparison. All in all, the lower body appears to be that of a striding biped, though it is suggested that naledi’s gait would not have been typical for either Homo or an australopith.11 The hands have a long thumb, useful in fine manipulation, but at the same time the finger bones are rather curved, suggesting strong grasping that fits with other elements of the upper body said to be indicative of climbing.12 Naledi then emerges as somewhat more australopith- than Homo-like in its body structure, but with some features typical of neither. The same could be said also of its cranial anatomy: unlike its australopith counterparts, the best-preserved cranium has a pronounced brow ridge reminiscent of some Homo, whereas brain volume is in the australopith range, and well below that of any other claimed Homo except those from Dmanisi. The teeth, on the other hand, are relatively small, and in this respect are more Homo- than australopith-like.

On balance, the preponderance of naledi’s anatomical attributes seem to align more with the australopiths than with any Homo (even the remarkably primitive supposed Homo from Dmanisi). Of course, the describers of the new species have concluded otherwise, which is why they opted to allocate their undoubted new species to Homo. But such assessments of general degrees of similarity are beside the point. The main lesson naledi has to teach us is the futility of trying to divide what is now a very extensive hominid record between australopiths and Homo. There is a lot more systematic structure—a more complex geometry of species relationships—in this record than can usefully be characterized by bundling the very diverse assemblage of hominids now known into one of two predetermined genera.

That much should be clear, yet the shock inflicted by Mayr on paleoanthropology back in 1950 still resonates in a general feeling among paleoanthropologists that it is somehow unbiological to allow hominid genera to multiply, as their early twentieth-century predecessors had heedlessly done. It is this residual collective trauma that explains both the widespread paleoanthropological reluctance to match the number of recognized genera to the plain facts of morphological diversity, and the corresponding readiness to expand existing genera beyond any rational anatomical structure.

Yet if there is one lesson that we can very clearly derive from a mushrooming fossil record, it is that the history of the hominid family is very much like that of any other successful family of mammals, among which diversification has always been the rule. The history of the hominids has been one of vigorous evolutionary experimentation with the hominid potential. Numerous species, whose relationships can only be clarified by recognizing several genera, have been thrown out onto an ever-changing ecological stage to compete and to flourish, or face extinction.

Chance and contingency loom large in this process, which is radically different from the slow, steady slog from primitiveness to perfection envisaged by Mayr. And this is important, for Mayr’s gradualist model implies that we have somehow been closely molded by nature to be a particular kind of organism. It implies that many of our features and behaviors have been programmed into us by eons of natural selection, thereby relieving us of some degree of responsibility for how we interact with each other and with the world. If we can adopt a more realistic notion of what our genus Homo is, we will at the same time open the way not only to a better understanding of the process that produced us, but also to a more accurate perspective about the kind of creature we happen to be.

  1. Ernst Mayr, “Taxonomic Categories in Fossil Hominids,” Cold Spring Harbor Symposia on Quanitative Biology 15 (1950): 109–18. 
  2. Louis Leakey, Phillip Tobias, and John Napier, “A New Species of Genus Homo from Olduvai Gorge,” Nature 202, no. 4,927 (1964): 7–9. 
  3. Jean de Heinzelin et al., “Environment and Behavior of 2.5-Million-Year-Old Bouri Hominids,” Science 284, no. 5,414 (1999): 625–29; Sharon McPherron et al., “Evidence for Stone-tool-assisted Consumption of Animal Tissues Before 3.39 Million Years Ago at Dikika, Ethiopia,” Nature 466, no. 7,308 (2010): 857–60; Sonia Harmand et al. “3.3-million-year-old Stone Tools from Lomekwi 3, West Turkana, Kenya,” Nature 521 (2015): 310–15. 
  4. G. Philip Rightmire et al., “Anatomical Descriptions, Comparative Studies, and Evolutionary Significance of the Hominid Skulls from Dmanisi, Republic of Georgia,” Journal of Human Evolution 50, no. 2 (2006): 115–41. 
  5. Bernard Wood and Mark Collard, “The Human Genus,” Science 284, no. 5,411 (1999): 65–71. 
  6. Ian Tattersall, The Fossil Trail: How We Know What We Think We Know About Human Evolution, 2nd edn., (New York: Oxford University Press, 2008). 
  7. Brian Villmoare et al., “Early Homo at 2.8 Ma from Ledi-Geraru, Afar, Ethiopia,” Science 347, no. 6,228 (2015): 1,352–55. 
  8. Jeffrey Schwartz and Ian Tattersall, “Defining the genus Homo,” Science 349, no. 6,251 (2015): 931–32. 
  9. Peter Brown et al., “A New Small-Bodied Hominin from the Late Pleistocene of Flores, Indonesia,” Nature 431 (2004): 1,055–61. 
  10. Lee Berger et al., “Homo naledi, a New Species of the Genus Homo from the Dinaledi Chamber, South Africa,” eLife 4 (2015), doi:10.7554/eLife.09560. 
  11. William Harcourt-Smith et al., “The Foot of Homo naledi,” Nature Communications 6, no. 8,432 (2015), doi:10.1038/ncomms9432. 
  12. Tracy Kivell et al., “The Hand of Homo naledi,” Nature Communications 6, no. 8,431 (2015), doi:10.1038/ncomms9431.