Paleolithic

← Encyclopedia

The Paleolithic, or Old Stone Age, is the long opening stretch of human prehistory — the era of chipped stone tools that begins with the earliest toolmaking, more than two million years ago, and closes with the end of the last Ice Age and the turn to farming. Across the overwhelming majority of that span the record is mute and practical. A hand-axe knapped from a flint nodule; a ring of fire-cracked stones where a hearth once burned; the cut-marked, marrow-split bones of hunted horse, reindeer, and aurochs. These are the ordinary residues of a creature making a living from a cold and shifting world, and for most of the period they are nearly all there is. The deeper one reaches back — through the Acheulean tool industries of Homo erectus and earlier hominins to the first Oldowan flakes of the African savanna — the more the evidence narrows to the bare grammar of survival: striking, cutting, burning, eating, moving on.

Then, late in the sequence and concentrated in the last forty or fifty thousand years, something else surfaces. It does not announce itself. It is recovered from the floors of caves and the fill of graves, and it has drawn the attention of nearly everyone who has ever asked the question of when, in the human story, religion begins.

The three kinds of evidence

Three classes of find carry that question, and it is worth holding them apart, because each fails the next in a different way.

The first is the painted caves of Ice Age Europe. At Lascaux in the Dordogne, at Altamira on the Cantabrian coast, at Chauvet in the Ardèche gorge, and at scores of lesser sites, people went deep underground — past daylight, past the inhabited mouth of the cave, sometimes hundreds of meters into passages reachable only by crawling — and there, by the light of tallow lamps and torch, they drew. They drew bison and horses, aurochs and red deer, mammoth and woolly rhinoceros, and the great predators: cave lions, bears, the rare and unsettling human-animal hybrid. The animals are observed with a precision that took thirteen thousand years and the invention of photography to recover; the human figures, where they appear, are schematic, masked, sometimes part-beast. The dating is now firm and startling. Charcoal from the Chauvet drawings and the calcite that later sealed them place the cave’s two phases of human use at roughly 37,000 and 31,000 years ago, making its lions and rhinos older than the celebrated bestiaries of Lascaux and Altamira by a span longer than all of recorded history. Lascaux belongs to the late Magdalenian, around 17,000 years ago; Altamira’s polychrome ceiling spans the Gravettian to the mid-Magdalenian. And the practice is not European alone: on the Indonesian island of Sulawesi, a red-ochre warty pig at the cave of Leang Tedongnge has been dated by uranium-series methods to at least 45,500 years — older than anything yet found in Europe, and proof that the impulse to put a recognizable animal on a cave wall was already loose across the breadth of the inhabited Old World.

The second is deliberate burial. Long before any cave was painted, bodies were being placed in the ground with evident care — laid out, flexed or extended, sometimes oriented, and dusted or packed with red ochre, the iron-rich pigment that recurs across the whole record like a refrain. At Qafzeh in the Levant, anatomically modern dead were interred with ochre and, in one grave, the antlers of a fallow deer arranged over a child, roughly a hundred thousand years ago. Far later and far richer, the graves at Sungir near Vladimir, Russia, dated to around 34,000 years ago, hold an adult man and two children sewn into garments worked with many thousands of mammoth-ivory beads — beads whose manufacture has been estimated at thousands of hours of labor — together with ivory spears straightened from the curve of tusks, pierced fox canines, and carved discs, the whole reddened with ochre.

The third is the small carved female figures found across Eurasia from the Pyrenees to Siberia, conventionally and misleadingly called “Venus” figurines — a nineteenth-century label borrowed from a Roman goddess to whom they bear no relation and carry no demonstrable connection. (The name is a modern imposition, not a recovered identification, and it is best held at arm’s length.) The earliest, the ivory figure from Hohle Fels in the Swabian Jura, comes from the basal Aurignacian and is at least 35,000 years old — the oldest unambiguous depiction of a human being yet known. Most are palm-sized, headless or faceless, with breasts, belly, and hips emphasized and the limbs tapering away; some, like the limestone Venus of Willendorf, are corpulent, others slender; they were carried, handled, sometimes broken. From the same Swabian caves comes a stranger object still: the Lion Man of Hohlenstein-Stadel, a foot-tall figure of mammoth ivory with a human body and the head of a cave lion, carved in the Aurignacian some forty thousand years ago — the oldest known representation of a being that does not, and cannot, exist.

What cannot be recovered

What any of this meant cannot be recovered, and the honest statement of the problem is the first thing a serious account must make. There are no texts. The people who knapped the flints, reddened the dead, and crawled into the dark left no account of themselves; they could not, the technology of writing lying tens of thousands of years in their future. The words by which the question is even posed — religion, the sacred, the soul, ritual, magic — are later categories, every one of them assembled in literate societies with their own quarrels and pressed backward onto a silence that cannot answer for itself. They are tools for thinking about the evidence, not contents extracted from it.

The discipline that grew up around these finds has been careful, sometimes to the point of austerity, about the line between what is recovered and what is inferred. The ochre is real and datable. The grave goods are real and datable. The depth of a painted gallery can be measured to the meter. Whether ochre on a corpse implies a belief that the corpse goes somewhere, whether ivory beads sewn into a child’s clothing express a hope held about the dead rather than a status held in life, whether a bison eight hundred meters underground was a prayer, a lesson, a boast, or a thing seen behind closed eyes — none of that is recovered. It is the inference laid over the artifact, and the artifact, mute, supports several inferences at once.

This is sharpest for the caves. The readings have been many, and they have succeeded one another less by disproof than by fashion. The earliest twentieth-century syntheses, above all the Abbé Henri Breuil’s, read the animals as hunting magic — images made to seize, by likeness, power over the beasts that fed the band. The mid-century structuralism of André Leroi-Gourhan and Annette Laming-Emperaire dismantled the hunting reading and proposed instead a patterned symbolic order, the species distributed across cave space by a grammar of pairing and opposition, often glossed as male and female. Others have read the deep galleries as initiation settings, stages for the making of adults or of specialists, the difficulty of access being itself the point. A widely debated proposal of the late twentieth century, advanced by David Lewis-Williams and Thomas Dowson, reads the recurring geometric signs — grids, dots, zigzags, nested curves — as entoptic phenomena, the luminous forms that arise in the human visual system under altered states, and the deep cave as a place entered to seek visions, its walls a membrane against which the beings of trance were fixed. And a steady counter-tradition insists that some or all of this may be art with no use the modern eye would file under religion at all — observation, pleasure, teaching, play, the marking of place — and that to read worship into every painted horse is to find in the cave only what one carried into it.

Each reading has been argued with rigor; none commands the field; and the figurines fracture the same way. They have been called mother-goddesses and the icons of a primordial fertility cult; charms for safe childbirth; self-portraits made by women looking down at their own bodies, which would account for the foreshortened, headless proportions; tokens of exchange or alliance between distant groups; and, with as much warrant as any, toys, dolls, or simply objects whose function was never single. A carved woman from thirty thousand years ago can be made to say almost anything, and the more confidently a reading is stated, the more it tends to reveal of its reader.

The shaman question and the first dawn of the sacred

The trance reading deserves a separate word, because it is the door through which the Paleolithic enters esoteric and comparative thought. Lewis-Williams and his colleagues drew their model partly from the ethnography of living practitioners — the San of southern Africa, and, more loosely, the trance-and-spirit complexes of northern Eurasia generalized under the label shamanism. That label is itself a constructed one: the word šamán is genuinely Tungusic, from the Evenki of Siberia, but the universal “-ism” was assembled by Enlightenment travelers, Romantic theorists, and twentieth-century scholars out of heterogeneous local practices, and it flattens roles that their own peoples never treated as one thing. To read a thirty-thousand-year-old painted gallery through a category built in this way — and built, in much of its documentary base, by imperial and colonial observers — is to reach across an abyss of time and place with a borrowed and contested instrument. The neuropsychology of vision may well be a human constant; that the Ice Age painters used it as a Siberian or San practitioner is held to is a hypothesis, not a finding, and the most careful proponents say so.

It is from exactly this material that a larger and looser idea has been built: that the Paleolithic is the first dawn of the sacred, the headwater from which the entire history of religion descends in unbroken line — that the ochre and the painted bull and the carved goddess are the visible beginning of a single current running through the Neolithic temple, the Bronze Age sanctuary, and every later church and rite. This framing has been enormously productive for the imagination and is itself an interpretation — a story told about the evidence rather than read out of it, and a contested one. The gap between a pigment stain on a cave floor and a doctrine about the soul is not a gap the evidence has any means of closing. What the deep-time record offers is gesture without gloss; the continuous narrative is supplied afterward, by those who need the beginning to lead somewhere.

What the Paleolithic does not contain, it is worth saying, belongs to its sequel and not to it. The settled village, the worked field, the monument built for the dead and the unseen — these come with the turn to farming and define the Neolithic, with the post-glacial foragers of the Mesolithic standing between. The Old Stone Age ends where the ice retreats and the first seeds go into the ground; its question about beginnings does not end there at all.

Reading the record

There is a textual and scholarly tradition here as real as the painted walls, and it has its monuments. The Hohle Fels figurine was published by Nicholas Conard in Nature in 2009, resetting the date of the earliest human figurative art. The case for cave art beyond Europe rests on the uranium-series dating of the Sulawesi paintings, reported by Adam Brumm, Maxime Aubert, and colleagues in Science Advances in 2021. The high-precision chronology that fixed Chauvet’s two phases of use was established by Anita Quiles and a large team in the Proceedings of the National Academy of Sciences in 2016. The trance hypothesis received its canonical statement in Lewis-Williams and Dowson’s 1988 Current Anthropology paper on entoptic phenomena; the deep antiquity of ochre as a symbolic, non-utilitarian material was argued by Erella Hovers and colleagues from the Qafzeh graves in the same journal in 2003. Around these stand the classic syntheses — Breuil’s catalog of the painted caves, Leroi-Gourhan’s structural decoding of cave space and his successors’ steady insistence that the images form a system before they form a creed — and the running modern argument over whether any of it can be read at all. The literature is large; what it converges on is the size of the silence, not a way around it.

What can be said plainly is narrow and strange enough on its own. Long before writing, before cities, before any name that survives for a god, human beings were carrying their dead into the ground reddened with ochre and laden with months of labor, and crawling by firelight into passages no one lived in to fix the shapes of animals on stone, for motives no surviving word quite fits. This deep-time record is where the question of how the sacred began is posed most sharply and met most poorly: every category carried to it — religion, art, magic, the shamanic trance — was minted in a far later world, and fits the painted dark the way a borrowed coat fits a stranger.

→ Related: Indus Valley Civilization · Mesopotamia · Mesolithic · Neolithic · Ritual · Religion · Soul · Magic