Proto-writing Stone Age artifacts are forcing a rewrite of one of humanity’s oldest assumptions.
A study published in the Proceedings of the National Academy of Sciences on February 24, 2026 analysed more than 3,000 geometric signs carved into 260 objects from cave sites in the Swabian Jura, a mountain range in southwestern Germany.
The objects date between 34,000 and 45,000 years ago. Some are tools. Some are small ivory figurines: a mammoth, a lion-human hybrid, a human figure with outstretched arms known as the Adorant.
All of them are covered in marks. Repeated lines. Dots. Crosses. Notches arranged in rows.
Archaeologists have been staring at these marks for over a century. The question was always the same: decoration, or something more? The 2026 study finally has an answer.
Something more. Significantly, measurably, demonstrably more.
What the Researchers Actually Did
Christian Bentz, a linguist at Saarland University, and Ewa Dutkiewicz, an archaeologist at Berlin’s Museum of Prehistory and Early History, did not try to decipher the symbols. They were clear from the outset that decipherment may be impossible and was not the goal.
Instead they measured the statistical structure of the sign sequences. Using computational techniques drawn from quantitative linguistics, they calculated the entropy of the carvings: how much information each sequence could carry, how predictable its symbols were, how they compared to known writing systems.
They then compared those statistical fingerprints to two other systems: proto-cuneiform, the earliest known writing system which emerged in Mesopotamia around 3,000 BCE, and modern alphabetic writing.
The result, Bentz said, was one he could not believe at first. He sent a screenshot to his colleague.
The 40,000-year-old Paleolithic signs were clearly distinct from modern writing. Modern alphabets directly encode spoken language and have a high, varied information density. The Stone Age marks are nothing like that.
But their statistical profile was a near-perfect match for proto-cuneiform. A system that would not be invented for another 35,000 years.
What That Gap Means
Consider what 35,000 years represents.
Modern humans have had agriculture for roughly 12,000 years. The Roman Empire lasted about 500 years. The entire span of recorded history, from the first Mesopotamian clay tablets to today, covers approximately 5,000 years.
The gap between these Stone Age carvings and the proto-cuneiform they statistically resemble is seven times longer than all of recorded history.
And yet the structural properties of the two systems are, in Bentz’s word, indistinguishable.
This does not mean Stone Age hunters in southern Germany invented writing and then forgot about it. It does not mean there is an unbroken chain of transmission between a mammoth figurine in a German cave and a clay tablet in ancient Mesopotamia. The researchers are careful to say that the signs may never be decoded and that what they were used for remains genuinely unknown.
What it means is that the cognitive capacity to create structured symbolic systems with measurable information density did not arrive suddenly in Mesopotamia 5,000 years ago.
It was already present, already being used, already stable across 10,000 years of consistent application, deep in the Ice Age.
The history of human communication is not a straight line from silence to speech to writing. It is something far longer and more complicated.
The Rules Nobody Taught
There is a detail in the study that is easy to skip past and should not be.
Crosses appear frequently on animal figurines, mammoths, horses, and on tools. They appear across the entire 10,000-year span of the objects in the study.
Crosses never appear on human figurines. Not once. Across 260 objects and 3,000 signs, this pattern holds absolutely.
Dots never appear on tools. Again, not once.
These are not accidental absences. They are rules. Shared conventions that were observed consistently across generations, across communities, across millennia. Bentz described it as a kind of taboo or convention: there was clearly something that meant crosses should not go on people.
Other studies have suggested that crosses in this period could signify ritual killing. Perhaps that is why they were kept off human forms. The honest answer is that nobody knows.
But the fact that the rules existed, that they were transmitted and maintained for thousands of years without any known writing system to preserve them, is itself an extraordinary finding.
It means these people had culture in the full sense. Shared symbolic frameworks. Inherited conventions. Information that outlasted the individual who first created it.
The Adorant and the Mammoth
Two objects in the study are worth dwelling on.
The Adorant is a small ivory plate about 38,000 years old, found in Geißenklösterle Cave. One side depicts an anthropomorphic figure with outstretched arms. The back is covered in rows of notches and dots arranged in sequences. Dutkiewicz has suggested these may reflect calendric observations: tracking lunar cycles or the passage of seasons.
The mammoth figurine from Vogelherd Cave, approximately 40,000 years old, is covered in crosses and dots on its surface. It is carved from mammoth ivory. Someone took the tusk of the animal it depicts, carved the animal’s form from it, and then covered that form in structured symbols.
Both objects fit in the palm of a hand. Dutkiewicz noted this specifically. You are able to see that they were carried, she said. The objects moved with the people. Like a proto-cuneiform tablet, they were portable. Personal. Meant to be held and consulted and transported.
Forty thousand years ago, a person carried a small carved mammoth with symbols on it in their hand, the same way you carry your phone in your pocket. The information was different. The impulse was the same.
What Genevieve von Petzinger Already Knew
The Bentz and Dutkiewicz study did not arrive in a vacuum.
Palaeoanthropologist Genevieve von Petzinger has spent years cataloguing geometric symbols found in cave art across Europe. She identified 32 recurring geometric signs that appear across European cave sites spanning tens of thousands of years. The same signs. The same 32. Reproduced in caves from Spain to Russia by people who, by most accounts, had no direct contact with each other.
Von Petzinger welcomed the new findings and described the approach as two excellent methods for confirming these marks were meaningful beyond decorative doodles. She has argued for years that the origins of symbolic communication stretch back far further than the archaeological consensus has been willing to accept.
The PNAS study provides the quantitative backbone that anecdotal observation and catalogue work could not. It gives the field a statistical method for distinguishing structured symbolic systems from decoration, one that can be applied to other sites, other objects, other periods.
The Longer Story
Bentz offered a sentence that deserves to close any article about this study.
The human ability to encode information in signs and symbols was developed over many thousands of years. Writing is only one specific form in a long series of sign systems. We continue to develop new systems for encoding information. Encoding is also the basis of computer systems.
The marks on the mammoth figurine, the dots on the Adorant’s back, the crosses on the tools, and the characters on this screen are points on the same long line. The line did not begin in Mesopotamia. It began, at least, in a cave in southern Germany, 40,000 years ago, when someone picked up a piece of ivory and decided that marks meant something.
We are still doing exactly that. We just have not deciphered theirs yet.
Read next: The Psychology of the News Cycle: Why Certain Stories Hit Different
What is dharma? what the mahabharata knew about losing everything
