Quite a lot about what ancient people could do, and almost nothing about what they felt. As of September 2026, the oldest dated cave art is a faded hand stencil on Muna Island, Indonesia, made at least 67,800 years ago. The oldest dated painting of animals and human-like figures, a scene in Sulawesi, Indonesia, is at least 51,200 years old (Oktaviana et al. 2026; 2024). Paintings like these show people planning, working with pigments, and making images that stand for things. They do not tell us what the images meant, what their makers believed, or anything about anyone’s mental health.
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.
How do scientists date a painting on a cave wall?
The mineral pigment in red ochre is not itself organic carbon, so it cannot be radiocarbon dated directly (Pons-Branchu et al. 2026, §5.3). Instead, scientists look for the thin mineral crust that water leaves as it seeps over a cave wall. That crust contains tiny amounts of uranium, which decays at a known rate, so the crust can be dated. Because it formed on top of the paint, its age is a floor: the painting is at least that old, and it could be older (Oktaviana et al. 2026, Methods).
Analogy: picture a flyer stapled over a poster. If the flyer is dated March, you know the poster went up before March. You don’t know how long before.
There’s a catch. If water later washed uranium out of the crust, the date comes out too old, so dating teams test for this (Oktaviana et al. 2026, Methods). A 2026 study at Altamira, Spain, found some uranium dates that ran too old, most likely for this reason (Pons-Branchu et al. 2026).
That is why the record keeps moving. In 2019, a Sulawesi hunting scene at a cave called Leang Bulu’ Sipong 4 was dated to at least 43,900 years. A sharper laser version of the method, published in 2024, pushed that floor to at least 48,000 years (Oktaviana et al. 2024). [Documented] The art didn’t get older. The measurement got better.
Where is the oldest cave art?
Sulawesi and nearby islands, Indonesia. An Indonesian–Australian team, with first author Adhi Agus Oktaviana, dated a crust over a hand stencil in Liang Metanduno, a cave on Muna Island, to at least 67,800 years ago. What survives is a 14 × 10 cm patch of faded pigment. One fingertip looks deliberately narrowed, a style found so far only in Sulawesi. The same rock panel shows two rounds of painting at least 35,000 years apart (Oktaviana et al. 2026). [Documented, as a minimum age]
The oldest dated figurative cave scene reported by this team is nearby. At Leang Karampuang, a big red pig is shown with at least three human-like figures, painted at least 51,200 years ago (Oktaviana et al. 2024). At Leang Tedongnge, which a survey team led by archaeologist Basran Burhan found in 2017, a pig painting is at least 45,500 years old (Brumm et al. 2021). [Documented]
Europe. On December 18, 1994, three cave explorers, Jean-Marie Chauvet, Éliette Brunel, and Christian Hillaire, found the French cave that now carries his name. According to France’s Ministry of Culture, Brunel spotted the first painting the group noticed, a small red mammoth (Ministry of Culture, “The Discovery”). More than 350 dates show two separate periods of human visits, about 37,000–33,500 and 31,000–28,000 years ago. Rockfalls sealed the entrance by about 21,500 years ago (Quiles et al. 2016). [Documented] You may see Chauvet dated to “32,000–30,000 years ago.” That figure comes from the first radiocarbon results, reported before conversion into calendar years (Quiles et al. 2016).
Lascaux, in southwestern France, was found on September 12, 1940, by Marcel Ravidat and three other village boys, Jacques Marsal, Georges Agnel, and Simon Coencas (Ministry of Culture, Lascaux). New radiocarbon dates on bone left in the cave point to visits about 21,500–21,000 years ago. The same official site also still carries an older estimate of about 23,000 years (Ministry of Culture, “Dating the figures”). [Disputed] Either way, a date of around 15,000 BCE, which you may still see, is too young for both estimates.
Australia and Africa. In 2021, a team reported a kangaroo on Balanggarra Country in the Kimberley as Australia’s oldest securely dated painting, at 17,500–17,100 years old (Finch et al. 2021). Excavations at Madjedbebe suggest people reached northern Australia by about 65,000 years ago (Oktaviana et al. 2026), so older paintings may exist [Plausible]. In southern Africa, the earliest widely accepted image-making by our species is geometric patterns cut into pieces of ochre about 100,000–75,000 years ago (Oktaviana et al. 2024). At Blombos Cave in South Africa, researchers reported a cross-hatched pattern drawn with ochre on a stone flake about 73,000 years ago (CNRS 2018). Neither is a cave-wall painting, but both show that image-making was not a European invention.
So where did art begin? [Unknown] Even the Sulawesi team calls it an open question whether picture-making first arose in Africa or after people spread out of it (Oktaviana et al. 2024). What gets dated depends on what survived, what has been searched, and where mineral crusts happened to form.
Who made the oldest art?
For Muna, the dating team says plainly that they have no obvious method to tell which human species made the stencil, and that other, earlier human species lived on Sulawesi. They credit Homo sapiens because of the stencil’s deliberate finger shaping and because modern humans were in the region at about that time (Oktaviana et al. 2026). [Plausible]
In 2018, another team reported minimum ages above 64,000 years for paintings in three caves in Spain. Because modern humans had not yet reached Europe, they argued that Neanderthals made them (Hoffmann et al. 2018). Other researchers have challenged that dating (Oktaviana et al. 2024, Methods). [Disputed]
What were the paintings for?
Nobody knows. [Unknown] The history of guessing is still worth a look.
- Doubt. Around 1868, a local man, Modesto Cubillas, found the Altamira cave in northern Spain. In 1879, Marcelino Sanz de Sautuola returned with his daughter María, who first saw the painted ceiling. He published the paintings as prehistoric in 1880, but scholars doubted him for over 20 years, until the French prehistorian Émile Cartailhac published his “mea culpa of a skeptic” in 1902 (Spanish Ministry of Culture, historical preamble). [Documented] The age of Altamira’s oldest art is [Disputed]: the Ministry’s historical account reports a minimum uranium date of about 35,500 years for a red sign (2026 preamble), but a 2026 cross-check found uranium loss could make sampled crusts look too old; its oldest validated over-paint crust was about 22,800 years old. This does not date every painting in Altamira (Pons-Branchu et al. 2026).
- Hunting magic. Some bison in French caves have arrow-like marks on their sides, so early researchers linked the images to magic for the hunt. As more caves were found, such marked animals turned out to be rare (Ministry of Culture, “Interpretations”). At Chauvet, mammoths, big cats, rhinoceroses, and bears make up almost 65% of the identifiable animals, and these were generally not hunted (Clottes, “Aurignacian cave art”). France’s official Lascaux site notes that cave animals don’t reflect what people ate (Ministry of Culture, “Themes”). [Disputed]: these observations challenge hunting magic as a single explanation for the art.
- Hidden structure. In 1957, Annette Laming-Emperaire argued that the animals were combined on purpose, in male/female pairings. André Leroi-Gourhan built this into a statistical model linking motifs to their places in the cave (Ministry of Culture, “Interpretations”; “Laming-Emperaire”; “Leroi-Gourhan”). [Disputed]
- Altered states. In the late 1990s, Jean Clottes and David Lewis-Williams linked many cave images to shamanism (Ministry of Culture, “Interpretations”). Their 1996 book’s subtitle translates as “trance and magic in the decorated caves” (library record). [Disputed] A note of care: this is a debate about ancient images. It is not evidence that altered states are safe or helpful, and nothing here suggests trying them.
- Stories. The Sulawesi team reads some figures as part-human, part-animal beings, and suggests the scenes reflect storytelling, perhaps myth (Oktaviana et al. 2024). [Plausible]
No one has shown which idea, if any, fits a given wall.
So what does cave art tell us about ancient minds?
It shows skills directly. People sprayed pigment around hands (Oktaviana et al. 2026) and painted animals with brushes and fingertips (Brumm et al. 2021). At Chauvet they shaded animals to show their shape and tried out perspective (Clottes). [Documented] In Sulawesi they grouped figures into scenes (Oktaviana et al. 2024). On nearby Muna Island, one stencil’s fingertip appears to have been narrowed on purpose (Oktaviana et al. 2026).
The Sulawesi researchers discuss image-making and storytelling as inferences from the arrangements of figures (Oktaviana et al. 2024, Implications). [Plausible] A picture does not give direct access to its maker’s thoughts.
What cave art can’t show is an inner life. You may read that cave art “unlocked” human culture or rewired the brain. We found no evidence for either claim. Our rule for material evidence is to climb from what is observed to what it may mean, and to stop where the evidence stops. With cave art, that stopping point comes before beliefs, feelings, trance, diagnosis, or healing. [Unknown] Rock art is not a patient chart.
How we know
What evidence exists: pigment on cave walls; mineral crusts that give minimum ages (Oktaviana et al. 2026); charcoal drawings that can be radiocarbon dated, and dates from rockfalls (Quiles et al. 2016); dates from bones left in caves (Ministry of Culture, “Dating the figures”); and, in Australia, mud-wasp nests built under and over paintings (Finch et al. 2021). Almost every date is a range or a minimum.
Whose voices are missing: The painters left no words. Everything about meaning comes from people living tens of thousands of years later. In Australia, rock art is part of living cultures. Balanggarra Traditional Owners gave permission for the Kimberley dating work and asked that exact site locations not be published (Finch et al. 2021, Acknowledgements and Data availability). This page doesn’t present their knowledge of meaning, because it isn’t ours to share. For Sulawesi, the sources we read record researchers, not the views of communities who live near the caves. That is a gap.
How power shaped the record: Established experts doubted Altamira for over 20 years (Spanish Ministry of Culture, historical preamble). Chauvet received a state-funded team of specialists and is described as the best-dated cave-art site in the world (Ministry of Culture, “Research Programme”). Rock art in Southeast Sulawesi was first reported in 1977 and stayed under-studied for decades (Oktaviana et al. 2026). Theories built on Europe’s well-studied caves shaped what “early art” was expected to look like, until the Sulawesi dates challenged them (Oktaviana et al. 2024).
Then vs. now
| What people believed then | What current evidence shows |
|---|---|
| 1880–1902: Many experts doubted that Altamira’s paintings were prehistoric. | Accepted as prehistoric since 1902. [Documented] Its oldest art’s exact age is [Disputed]. |
| Ideas about early art were built mainly on Europe’s caves. | The oldest dated cave art is in Indonesia. Where art began is [Unknown]. |
| The paintings were hunting magic. | Challenged as a single explanation. [Disputed] |
| The paintings show shamans in trance. | [Disputed] |
| Early art was rarely about people or stories. | Sulawesi scenes of human-like figures are at least 51,200 years old. [Documented] |
| Cave art shows what ancient people felt, feared, or suffered. | It can’t. Beliefs, emotions, and mental health stay [Unknown]. |