Nobody knows exactly when. A thought leaves no fossil. What survives are objects that look like they were made to mean something: perforated sea-snail shells with traces of red pigment, probably worn on a cord, in Morocco about 82,000 years ago; a cross-hatched drawing on a stone flake in South Africa about 73,000 years ago; ostrich-eggshell water flasks engraved with repeated patterns about 60,000 years ago; a painted scene of human-like figures and a pig in Sulawesi, Indonesia, at least 51,200 years ago (Oktaviana et al. 2024). Scholarly reviews read this record as a long, uneven buildup that shows up early and richly in Africa, not a sudden switch in Europe around 40,000 years ago (d’Errico & Stringer 2011; Dapschauskas et al. 2022). How it happened, why, and whether Neanderthals shared it are still live arguments.
How to read the labels on this page: [Documented] means the evidence shows it directly. [Strongly supported] means strong sources agree. [Plausible] means the evidence allows it but doesn’t prove it. [Disputed] means qualified experts disagree. [Unknown] means the evidence can’t answer the question.
What is a symbol, anyway?
A symbol is something that stands for something else because people agree it does.
Analogy (not evidence): a thumbs-up emoji is a few pixels until a group agrees it means “sounds good.” That’s the archaeologist’s problem: the agreement lived in people’s minds, and those people are long gone.
Is there a clean wall between human symbols and animal signals? In 2018, biologist Toshitaka Suzuki reported that Japanese tits, small songbirds, give one alarm call only for snakes. Birds hearing it approached a stick wiggled like a snake; after a general alarm call, they almost never did. Suzuki reads this as the call bringing a snake’s image to mind (Suzuki 2018). That doesn’t make the call a word. It does make “animals only signal, humans symbolize” too simple. [Plausible — our reading]
Why can’t we just dig up a thought?
Because each step from “object” to “meaning” is a separate claim that needs its own evidence. Let’s run one real find up the ladder.
Evidence lab: the Taforalt shells. At Grotte des Pigeons (Taforalt), Morocco, a team led by archaeologist Abdeljalil Bouzouggar of Morocco’s national archaeology institute (INSAP) reported 13 shells of a small sea snail, Nassarius gibbosulus, from layers their age model places between about 73,400 and 91,500 years ago (Bouzouggar et al. 2007).
- What’s there? Microscopic red iron-oxide residue on ten shells (nine perforated, one not), and a shiny, scratched wear pattern on ten. [Documented]
- Natural or human? Only 3.5% of dead shells in a modern comparison sample have that hole position, and the cave sat about 40 km from the sea: people chose these shells and carried them inland. The authors think people made the holes too, but can’t firmly say. [Strongly supported for selection; Plausible for human-made holes]
- Used how? The authors call the wear possible evidence that the shells hung on a cord. [Plausible]
- Did they mean something? Shell beads like these are widely accepted by archaeologists as signs of symbolic behavior (Hoffmann et al. 2018). The usual argument: ornaments tell other people something about the wearer. [Plausible to Strongly supported — a widely accepted inference, still an inference]
- Meant what? Belonging? Status? A gift? Beauty? [Unknown]
The authors add a brake: the oldest ornament found doesn’t show when the ability to make one evolved.
What are the oldest clues?
Many cave-art dates come from thin mineral crusts that formed over the paint, so they are minimum ages: the art could be older (Oktaviana et al. 2024).
| About when | Where | What was found | Label |
|---|---|---|---|
| About 320,000 years ago | Olorgesailie, Kenya | Red ochre and black manganese collected and processed; obsidian brought from distant sources (Smithsonian research team 2018) | Find [Strongly supported]; symbolic use [Unknown] |
| From about 160,000 years ago | Africa (review of 100+ sites) | Ochre use becomes habitual (Dapschauskas et al. 2022) | Pattern [Strongly supported]; ritual meaning [Disputed] |
| 120,000–115,000 years ago | Cueva de los Aviones, Spain | Perforated, pigment-stained shells in a Neanderthal-associated site (Hoffmann et al. 2018) | Dates [Documented]; symbolic reading [Disputed] |
| About 100,000 years ago | Skhul, Israel; Blombos Cave, South Africa | Perforated shells; abalone shells used as ochre-paint containers (d’Errico & Stringer 2011; Henshilwood & van Niekerk 2018) | [Strongly supported] |
| About 78,000 years ago | Panga ya Saidi, Kenya | A child of about two and a half to three, buried in a pit and quickly covered (Max Planck Society 2021) | Burial [Strongly supported]; beliefs [Unknown] |
| 75,000–73,000 years ago | Blombos Cave, South Africa | 49 shell beads; a cross-hatched line pattern drawn with ochre on a stone flake (d’Errico & Stringer 2011; Henshilwood & van Niekerk 2018) | [Strongly supported] |
| At least 67,800 years ago | Muna Island, Indonesia | Faded hand stencil (Oktaviana et al. 2026) | Minimum age [Documented]; maker [Plausible: Homo sapiens] |
| 65,000–55,000 years ago | Diepkloof, South Africa | 270 pieces of engraved eggshell from at least 25 flasks (Texier et al. 2010) | [Documented] |
| At least 51,200 years ago | Leang Karampuang, Sulawesi | Human-like figures with a pig (Oktaviana et al. 2024) | [Documented]; story [Unknown] |
| At least 35,000 years ago | Hohle Fels, Germany | Carved mammoth-ivory female figurine (University of Tübingen 2009) | Age [Strongly supported]; meaning [Unknown] |
| About 21,000–21,500 years ago | Lascaux, France | Painted cave (French Ministry of Culture) | Dating reported by the Ministry [Strongly supported]; date of every painting [Unknown] |
The 2011 review describes a flicker: shell beads seem to vanish from Africa and the Levant (the eastern Mediterranean coast) between about 70,000 and 40,000 years ago, then reappear almost everywhere in Africa and Eurasia (d’Errico & Stringer 2011).
“Firsts” move, too: a Sulawesi pig painting at least 45,500 years old, hailed in 2021 as the earliest known representational art (Brumm et al. 2021), was beaten by a nearby scene in 2024.
Sudden revolution or slow build?
Archaeologist Francesco d’Errico and paleoanthropologist Chris Stringer laid out three main stories in 2011 (d’Errico & Stringer 2011):
- The switch. A genetic change affecting the brain, about 50,000 years ago in Africa, unlocked modern behavior in people who already looked like us. D’Errico and Stringer cite archaeologist Richard Klein for this view. (This is d’Errico and Stringer’s summary of Klein’s position.)
- The slow build. In 2000, Sally McBrearty and Alison Brooks argued that the pieces of the so-called “human revolution” appear tens of thousands of years earlier in Africa’s Middle Stone Age, scattered widely in space and time. D’Errico and Stringer’s full review discusses this gradual African model (2011, Introduction).
- The flicker. New behaviors appear and disappear in Africa and Eurasia (Europe and Asia) between about 200,000 and 40,000 years ago, not only in our species, before they stick.
Where it stands: d’Errico and Stringer conclude that the evidence contradicts a sudden change in human thinking 40,000–50,000 years ago, in Europe or in Africa. [Strongly supported against a single sudden European switch; how it actually happened is Disputed]
Why would symbols flicker on and off?
One idea: bigger, better-connected groups keep skills alive; small, scattered groups lose them. Adam Powell and colleagues modeled this in 2009 (as summarized by d’Errico & Stringer 2011), and philosopher Kim Sterelny argued that behavior looks “modern” once groups can reliably pass know-how to the next generation (Sterelny 2011). In 2016, Krist Vaesen and colleagues argued these population models only work under implausible conditions and clash with the archaeological record (Vaesen et al. 2016). [Disputed]
What about the brain?
A study of 20 early Homo sapiens fossil braincases found that by about 300,000 years ago, at Jebel Irhoud, Morocco, our species’ brains were already modern in size. The rounder modern shape arrived gradually, reaching today’s range between about 100,000 and 35,000 years ago (Neubauer, Hublin & Gunz 2018). [Documented measurements] That roughly parallels the archaeology, but overlap isn’t cause, and a skull’s shape isn’t a thought. Did symbols reshape the brain, or the brain enable symbols? [Unknown]
Evolutionary psychologist Robin Dunbar’s “social brain hypothesis,” set out in a 1998 review, says primate brains grew to handle larger, more complex social groups, and he ties it to understanding other minds and to language (Dunbar 1998). It doesn’t argue about symbols directly; extending it to symbolic thinking is a further step. [Plausible at most]
Did Neanderthals think in symbols?
[Disputed] The Aviones shells and pigments, dated to 115,000–120,000 years ago, led their excavators to suggest the roots of symbolic material culture may go back to the common ancestor of Neanderthals and our species (Hoffmann et al. 2018). A 2018 study dating marks in three Spanish caves to over 64,800 years, and so to Neanderthals, drew a 2020 reply from Randall White and dozens of rock-art specialists arguing the thin mineral crusts dated can give misleading ages (White et al. 2020). D’Errico and Stringer argue that late Neanderthals in western Europe were fully able to adopt new kinds of symbolic objects (d’Errico & Stringer 2011). Experts still disagree.
And language?
[Unknown] No bead or drawing records speech, so which came first, or whether they grew together, is still an open question.
How we know
What exists: objects, residues, microscopic wear, a few burials, paintings, and dates. Many dates come from the surrounding layer or a crust over paint, not the thing itself (Lascaux’s come from charcoal, bone and antler left in the cave, French Ministry of Culture).
Whose voices are missing: every maker’s. We don’t know who strung a given bead or what anyone felt. Minds in distress are invisible to a trowel. Communities living near these sites today hold their own heritage. We don’t use later southern African knowledge to decode 73,000-year-old marks, or present sacred or culturally controlled knowledge. Aboriginal Australian rock art, a living tradition governed by its custodians, is not covered here.
How power shaped the record: The African evidence challenges a history built mainly from European finds (d’Errico & Stringer 2011). When research reaches new places, with researchers and institutions in Morocco, Kenya, and Indonesia among the leads and co-authors, the “firsts” move.
Then vs. now
Then: what the objects show people did. Gathered and processed pigments. Collected perforated shells, some bearing red pigment, and probably strung them. Engraved repeated patterns on water flasks. Drew lines with ochre. Buried a small child in a pit. Painted animals and human-like figures. What they believed while doing it: [Unknown].
How scientists told it in the late 20th century: a “human revolution” around 40,000 years ago, read mainly from Europe’s record.
Current evidence: many pieces show up earlier and unevenly: early and richly in Africa, but also in the Levant, Neanderthal Iberia (disputed), and Southeast Asia. No agreed single switch; even the word “modern” is contested.
Why this matters for a history of the mind: here we first have to say plainly, “we can see what hands did, not what minds felt.” Later chapters on moods, suffering, and healing start from that same honesty.