Writing system possibly identified in 40,000 year old artifacts

[Micael]

Very interesting. That it appears to have a complexity similar to proto-cuneiform tens of thousands of years earlier is mind boggling.

Scientists discover oldest form of writing in mysterious Stone Age engravings

By Ben Cost
Published Feb. 24, 2026, 10:32 a.m. ET
Ancient Sumerian writing.
Stone Age artifacts discovered in a German cave could push back the origins of writing by 30,000 years.
willbrasil - stock.adobe.com
The origins of writing aren’t set in stone.

The ancient cave peoples weren’t as illiterate as portrayed in popular media. Archaeologists have discovered Paleolithic glyphs in a German cave that could potentially push back the history of written communication by over 30,000 years, per a rock-solid study in the journal Proceedings Of The National Academy of Sciences.

According to the researchers, the symbols were engraved on artifacts that dated back some 40,000 years to the Stone Age, when early humans arrived in Europe from Africa and encountered the Neanderthals.

Despite their age, these ancient etchings boasted a complexity comparable to the early stages of the world’s oldest writing system, cuneiform, which originated around 5,000 years ago, the New Scientist reported.

“The artifacts date back to tens of thousands of years before the first writing systems,” exclaimed study co-author Ewa Dutkiewicz, an archaeologist at Berlin’s Museum of Prehistory and Early History, Popular Science reported.


Dutkiewicz and her team had came upon this writing revelation while investigating 260 relics discovered in cave repositories in the Swabian Jura, a mountainous region in Southwest Germany. This archaeological treasure trove included flutes, carvings of animals like mammoths, and figurines of animal-human hybrids.

They were etched with a total of 22 different recurring symbols, including a V-shaped notch and lines, crosses and dots.

Hoping to shed light on the symbols, the team inputted 3,000 of the inscriptions into a Stone Age sign database with the goal of seeing how they stacked up against later writing systems. They specifically compared their patterns to proto-cuneiform, the earliest form of pre-writing, which was engraved in clay in Mesopotamia circa 3500 to 3350 BC.


“It makes sense to look at sequences, because information is not only encoded in the number of different signs you have, but… in how you combine the signs,” said study collaborator, Christian Bentz a linguist at Saarland University in Saarbrücken. He noted that up until now, there had been very “little empirical work carried out on the basic, measurable characteristics of the signs.”

While these patterns differed drastically from modern writing, their configuration was “very similar” to their Mesopotamian counterpart, according to Bentz.

This suggested that the earliest H. sapiens in Europe, a group of hunter-gatherers, had devised a system to put their thoughts on paper, er stone — one of the hallmarks of writing — and that this system was as advanced as their descendants. In other words, their mode of symbolic thought transcription wasn’t restricted to making paintings on cave walls.

[Sukhoiman]

It is very interesting find for sure, but NYpost (and others) made the headline clickbaity/sensationalist well past the original authors:

https://www.pnas.org/doi/10.1073/pnas.2 ... i9_Q#sec-3

i.e not a writing system, but "conventional signs" that dont meet the criteria of writing (I remember seeing this w.r.t symbols in polynesia too somewhere)...but do bear resemblance to how proto-cuneiform shaped up in fertile crescent.

The key difference being latter's actual progression into (first we know of) a writing system: cuneiform.

I think the large factor at play for this progression to happen is the neolithic revolution and its particular intensity at hand in fertile crescent (and then how it spread/independently originated w.r.t Nile, Indus etc)....otherwise the capacity exists (since there are lot of arguments now that modern human cognitively speaking can be traced back to 100,000 years which is full magnitude of time compared to say mid 20th century understanding etc)....but actual writing development needs that food surplus consistently unlocked for dense enough sedentary civ and actual writing etc.

[Micael]

I think that depends on what criteria you put up for something to be a ”writing system.” In this case we have a graphical symbol system which encodes information, and was used by apparently a number of individuals over a very long time period. That points to standardization of said system as well. While it does not have all the features of a fully developed writing system such as that seen in cuneiform later on, I still think that this should qualify as a ”writing system.”

[Sukhoiman]

Just pointing out whats in the paper itself (https://www.pnas.org/doi/10.1073/pnas.2520385123), i.e they prefer "sign systems" for it :

"Our results strongly contradict the hypothesis that the sign sequences of the Swabian Aurignacian constitute writing in this strict sense. Their statistical properties are very different from those of genuine writing systems around the world. "

"In sum, the sign sequences of the Swabian Aurignacian might be described as human intercommunication by means of conventional visible marks, while they certainly do not meet the criterion of writing sensu strictu."

"Our analyses hence suggest that the first hunter-gatherers arriving in Central Europe more than 40,000 y ago already had the information capacity to create a sign system comparable to protocuneiform in terms of information encoding potential."

"It remains hard—or impossible—to prove that Aurignacian sign systems served the same numero-ideographic functions as protocuneiform. Moreover, there is another stark contrast between them: Protocuneiform developed into a full-blown writing system representing the Sumerian language within the subsequent 1,000 y. The sign sequences of the Swabian Aurignacian, on the other hand, were stable in terms of information density—for 10,000 y—and then disappear."

====================================================

I mean they omit Indus script for a reason from:

"Writing in the strict sense—i.e. sign sequences structurally representing spoken language—has developed independently at least three to four times, namely, in Mesopotamia, Egypt, China, and Mesoamerica (68), and potentially a fifth time in the Eastern Pacific (31). These developments fall in the range of roughly 2500 BC to 1500 AD."

.....as Indus script, be as it may during the neolithic revolution --->bronze age... doesnt have the length of signs to properly deduce its a writing system.

[Nik_SpeakerToCats]

Looks like 'Tally-man' markings, as contrived by herders and cricket scorers since antiquity...

[Belushi TD]

Looks like 'Tally-man' markings, as contrived by herders and cricket scorers since antiquity...

I was thinking something similar. One of the things that could possibly have been counted would be numbered in roughly factors of 28. Doing a quick count it appears that some of the lines are 13 and some are 14.

Belushi TD

[Micael]

Looks like 'Tally-man' markings, as contrived by herders and cricket scorers since antiquity...

I was thinking something similar. One of the things that could possibly have been counted would be numbered in roughly factors of 28. Doing a quick count it appears that some of the lines are 13 and some are 14.

Belushi TD

A lunar month has been one suggestion, which is about that lenght depending on how you count it. So the system could have for instance had a numerical base derived from that. Either it was a pure calendar setup that only recorded information as it related to the date, like on so-and-so date is the equinox, or it could have recorded other numerical information as well but in a 14/28 base system derived from the lunar month.

[Belushi TD]

I was actually aiming at a woman's menstrual cycle, rather than a lunar month.

The lunar month is pretty firm. The menstrual cycle can vary significantly.

Belushi TD