6. Archaic and transitional hominins

Aside, Card, CardGrid, LinkCard, Steps, Tabs, TabItem, Badge, } from '@astrojs/starlight/components';

Core idea

Between roughly 4 and 2 million years ago, Africa is full of upright apes. They are unambiguously hominins — closer to us than to chimpanzees — but they are not yet Homo. Their brains are only modestly larger than a chimpanzee's, their bodies are small, and their jaws and teeth still do most of the heavy lifting that modern humans now hand off to tools and fire. Wood calls them archaic hominins when they are clearly australopith-grade, and transitional hominins when their anatomy starts to drift toward our genus without fully arriving.

This is the messy middle of the family tree. Bipedalism is settled. What is not settled is the species count, the ancestor-descendant lines, or where the boundary of "Homo" should be drawn.

Wood's framing: Archaic hominins share more morphology with us than with chimpanzees, yet they lack the jaw, tooth, body, and brain changes that mark Homo. The transitional forms blur that line on purpose — the data refuses to be tidy.

Why it matters

If you only remember one ancestor from this window, it is Australopithecus afarensis — Lucy. But Lucy is a single point on a bush, not a step on a ladder. The 4-2 Mya record contains at least two distinct evolutionary experiments: a gracile lineage that keeps a generalist diet and modest body, and a robust lineage (the paranthropines) that doubles down on enormous chewing teeth and powerful jaw muscles. One of these experiments produces Homo; the other produces a dead end that nevertheless thrived for a million years.

Getting the geometry right matters because it reframes our origin story. We were not "destined" to become Homo. We were one branch of a successful African radiation — and for most of that window, the Paranthropus strategy looked at least as promising.

What changes when you see this period as a bush

A ladder view asks: what came before Homo? A bush view asks: which of several contemporary hominins is most likely our ancestor, and which are cousins that died out? Once you accept that afarensis, africanus, garhi, boisei, robustus, and the early Homo fossils overlap in time, the question stops being chronological and becomes anatomical: who shares the most derived features with later Homo?

Why the habilis problem is uncomfortable

If Homo habilis turns out to be more australopith than Homo in its body proportions and tooth size, the genus boundary has to move — either habilis gets demoted, or "Homo" has to swallow a creature that doesn't look much like us. Either move has consequences for how the rest of the tree is drawn.

Key takeaways

Mental model — the 4-2 Mya bush

Copy sourceZoom

The three branches in detail

Gracile australopiths — the generalist baseline

The gracile australopiths are the chassis on which everything else is built. Au. anamensis, from Kanapoi in Kenya (4.2-3.9 Mya), has chimp-like canines but the thick-enamelled chewing teeth that mark every later hominin — a hint that the dietary shift toward harder, more abrasive foods was already underway. Half a million years later, Au. afarensis dominates the record: two well-preserved skulls, dozens of jaws, and the partial skeleton known as Lucy from Hadar. A 3.3 Mya infant skeleton from Dikika (described by Zeray Alemseged) is even older than Lucy and confirms the body plan.

The portrait: 35-55 kg, 3-4 feet standing height, 400-500 cc brain (a touch larger than a chimpanzee's, but barely so once body size is controlled for), small incisors, large molars. The pelvis and lower limbs support bipedal walking but probably not long-distance walking; the upper limbs say the animal still climbed competently. The 3.66 Mya Laetoli footprint trails — preserved when a hominin family crossed wet volcanic ash that then set like concrete — close the case: bipedalism is real, habitual, and ancient.

Au. africanus, the southern African gracile (Taung, Sterkfontein, Makapansgat), is essentially the same grade of organism. Au. garhi (2.5 Mya, Bouri) is the odd one out: gracile body, but chewing teeth almost as big as a robust paranthropine's, and animal bones nearby that show cut marks from sharp-edged stone tools. Garhi may not be our ancestor, but it is a strong hint that somebody in this window was making tools.

Robust paranthropines — the heavy-chewing experiment

The robust paranthropines split off the gracile stem around 2.7 Mya. Paranthropus aethiopicus (the "Black Skull" from West Turkana) is the earliest. P. boisei dominates east Africa from 2.3 to 1.3 Mya; P. robustus runs in parallel in southern Africa, known from the cave breccias of Swartkrans, Kromdraai, and Drimolen. The shared feature is dramatic: gigantic chewing teeth, deep mandibles, flaring cheekbones to anchor huge temporalis muscles, and — in males of P. boisei — a sagittal crest along the top of the skull where those muscles attached.

This is not a failure. The paranthropines lasted about 1.5 million years — longer than Homo sapiens has so far existed. Their strategy was to specialise on hard, fibrous, or gritty plant foods that lesser jaws could not process. The strategy worked until it didn't: when the foods they were built for retreated, they had no alternative. They went extinct around 1 Mya.

Early Homo and the habilis problem

The earliest fossils assigned to our genus — Homo habilis and Homo rudolfensis, from East African sites between roughly 2.4 and 1.6 Mya — are where the bush gets genuinely uncomfortable. Their brains are slightly larger than an australopith's (around 510-700 cc), and their chewing teeth are slightly smaller. But their body proportions, when we have them, are still closer to afarensis than to later Homo. Long arms, short legs, small stature.

Wood's question, which he has pressed for decades, is whether habilis belongs in Homo at all. If the genus is defined by a package of features — encephalisation, smaller dentition, modern body proportions, full commitment to terrestrial bipedalism, prolonged development — then habilis clears maybe two of those bars. Depending on which definition you prefer, habilis is either the earliest Homo or the latest australopith with a slightly bigger brain.

The honest answer is that habilis is transitional — neither fully australopith nor fully Homo, exactly what we would expect to find sitting between the two grades. The category problem is ours, not the fossil's.

Practical application

How to read a "new ancestor" headline

When to be sceptical of a new species name

Example: reading Australopithecus sediba the parsimonious way

In 2010, two associated skeletons from the Malapa cave in South Africa (around 2 Mya) were named Australopithecus sediba and described as a mosaic of africanus and early Homo features — and proposed as a candidate ancestor of our genus. The claim is exciting but also exactly the kind of headline the diagnostic above is built for.

Apply the checklist. Where on the bush? Late gracile australopith, geographically southern, contemporaneous with both robustus and the earliest Homo in east Africa. Brain size? Around 420 cc — australopith range. Chewing teeth? Smaller than robustus, comparable to africanus. Body proportions? Mixed. Is a new species required, or could the Malapa fossils be a sub-population of africanus showing more variation than we previously sampled?

Wood's instinct — apply parsimony, treat sediba as a southern africanus variant until proven otherwise — is not a denial that the find is important. It is a refusal to multiply species when one will do. The same caution applies to Au. deyiremeda, Au. bahrelghazali, and any future "new ancestor" announcement built on a handful of fragments. The bush is bushy enough; we should be careful before adding more twigs.

Caveats

The Oldowan stone tool industry — simple flakes and choppers struck from cobbles — appears in the record around 2.6 Mya at sites like Gona in Ethiopia. Wood treats it as the marker that opens the transitional window. But who made the earliest Oldowan tools is genuinely uncertain: Au. garhi, early Homo, and even Paranthropus have all been proposed as candidate knappers. Tool-making is not a clean species marker.

Related material