© Dozier Marc/Photolibrary
Time to look around the paintings
The first intrepid explorers to brave the 7-metre crawl through a perilously narrow tunnel leading to the Chauvet caves in southern France were rewarded with magnificent artwork to rival any modern composition. Stretching a full 3 metres in height, the paintings depict a troupe of majestic horses in deep colours, above a pair of boisterous rhinos in the midst of a fight. To the left, they found the beautiful rendering of a herd of prehistoric cows. "The horse heads just seem to leap out of the wall towards you," says Jean Clottes, former director of scientific research at the caves and one of the few people to see the paintings with his own eyes.

When faced with such spectacular beauty, who could blame the visiting anthropologists for largely ignoring the modest semicircles, lines and zigzags also marked on the walls? Yet dismissing them has proved to be something of a mistake. The latest research has shown that, far from being doodles, the marks are in fact highly symbolic, forming a written "code" that was familiar to all of the prehistoric tribes around France and possibly beyond. Indeed, these unprepossessing shapes may be just as remarkable as the paintings of trotting horses and tussling rhinos, providing a snapshot into humankind's first steps towards symbolism and writing.

Until now, the accepted view has been that our ancestors underwent a "creative explosion" around 30,000 to 40,000 years ago, when they suddenly began to think abstractly and create rock art. This idea is supported by the plethora of stunning cave paintings, like those at Chauvet, which started to proliferate across Europe around this time. Writing, on the other hand, appeared to come much later, with the earliest records of a pictographic writing system dating back to just 5000 years ago.

Few researchers, though, had given any serious thought to the relatively small and inconspicuous marks around the cave paintings. The evidence of humanity's early creativity, they thought, was clearly in the elaborate drawings.

While some scholars like Clottes had recorded the presence of cave signs at individual sites, Genevieve von Petzinger, then a student at the University of Victoria in British Columbia, Canada, was surprised to find that no one had brought all these records together to compare signs from different caves. And so, under the supervision of April Nowell, also at the University of Victoria, she devised an ambitious masters project. She compiled a comprehensive database of all recorded cave signs from 146 sites in France, covering 25,000 years of prehistory from 35,000 to 10,000 years ago.

What emerged was startling: 26 signs, all drawn in the same style, appeared again and again at numerous sites. Admittedly, some of the symbols are pretty basic, like straight lines, circles and triangles, but the fact that many of the more complex designs also appeared in several places hinted to von Petzinger and Nowell that they were meaningful - perhaps even the seeds of written communication.

© New Scientist
Stone Age jottings
A closer look confirmed their suspicions. When von Petzinger went back to some of the records of the cave walls, she noticed other, less abstract signs that appeared to represent a single part of a larger figure - like the tusks of a mammoth without an accompanying body. This feature, known as synecdoche, is common in the known pictographic languages. To von Petzinger and Nowell, it demonstrated that our ancestors were indeed considering how to represent ideas symbolically rather than realistically, eventually leading to the abstract symbols that were the basis of the original study.

"It was a way of communicating information in a concise way," says Nowell. "For example, the mammoth tusks may have simply represented a mammoth, or a mammoth hunt, or something that has nothing to do with a literal interpretation of mammoths." Other common forms of synecdoche include two concentric circles or triangles (used as eyes in horse and bison paintings), ibex horns and the hump of a mammoth. The claviform figure - which looks somewhat like a numeral 1 - may even be a stylised form of the female figure, she says.

The real clincher came with the observation that certain signs appear repeatedly in pairs. Negative hands and dots tend to be one of the most frequent pairings, for example, especially during a warm climate period known as the Gravettian (28,000 to 22,000 years ago). One site called Les Trois-Frères in the French Pyrenees, even shows four sign types grouped together: negative hands, dots, finger fluting and thumb stencils (a rare subcategory of the negative hands).

Grouping is typically seen in early pictographic languages - the combined symbols representing a new concept - and the researchers suspect that prehistoric Europeans had established a similar system. "The consistency of the pairings indicate that they could really have had a meaning," says Nowell. "We are perhaps seeing the first glimpses of a rudimentary language system."

Lines, dots and love hearts

Von Petzinger caused quite a stir when she presented her preliminary findings last April at the Paleoanthropology Society Meeting in Chicago. She and Nowell have recently submitted a paper to the journal Antiquity and they are currently preparing another paper for the Journal of Human Evolution. The Smithsonian Institution's National Museum of Natural History in Washington DC plans to include the symbols in a forthcoming exhibition on human evolution.

"This work is really exciting," says Iain Davidson, an Australian rock art specialist at the University of New England in New South Wales. "We can see that these people had a similar convention for representing something."

© New Scientist
A group of 26 symbols crops up at Stone Age sites throughout the world – are these the origin of the written word?
Suspecting that this was just the beginning of what the symbols could tell us about prehistoric culture, von Petzinger and Nowell's next move was to track where and when they emerged. The line turned out to be the most popular, being present at 70 per cent of the sites and appearing across all time periods, from 30,000 to 10,000 years ago.

The next most prolific signs were the open angle symbol and the dots, both appearing at 42 per cent of the sites throughout this period. The vast majority of the remaining symbols are each present in around one-fifth of the French caves, the exceptions being the cordiform (roughly a love-heart shape), reniform (kidney shape), scalariform (ladder shape) and spiral, which all turned up in just a handful of sites. "The spiral only appears in two out of the 146 sites throughout the entire time period, which really surprised me as it is a common motif in many later cultures," says von Petzinger.

The Rhone valley and the Dordogne and Lot regions in the south seem to have been the original sites for the symbols in France: most signs seem to appear in these regions before spreading across the rest of the country. Notable exceptions include the zigzag, which first appeared in Provence and is a relative latecomer, debuting around 20,000 years ago.

No signs ever emerged in northern France, though. "For large periods of time the north was uninhabitable because of ice sheets coming and going, so there was less opportunity for culture to develop independently up there," says von Petzinger.

The Ice Age may have hindered the cultural revolution in the north, but elsewhere it could have been instrumental in furthering it. "People were forced to move south and congregate in 'refugia' during the last glacial maximum, 18,000 to 21,000 years ago, and it is at this time when we start to see an explosion in rock art," says Nowell. "One possibility is that they were using the signs to demarcate their territories."

Yet while long winters spent in caves might have induced people to spend time painting wonder walls, there are reasons to think the symbols originated much earlier on. One of the most intriguing facts to emerge from von Petzinger's work is that more than three-quarters of the symbols were present in the very earliest sites, from over 30,000 years ago.

"I was really surprised to discover this," says von Petzinger. If the creative explosion occurred 30,000 to 40,000 years ago, she would have expected to see evidence of symbols being invented and discarded at this early stage, with a long period of time passing before a recognisable system emerged. Instead, it appears that by 30,000 years ago a set of symbols was already well established.

Rewriting prehistory

That suggests we might need to rethink our ideas about prehistoric people, von Petzinger says. "This incredible diversity and continuity of use suggests that the symbolic revolution may have occurred before the arrival of the first modern humans in Europe." If she is right, it would push back the date of the creative explosion by tens of thousands of years.

The idea would seem to fit with a few tantalising finds that have emerged from Africa and the Middle East over recent years. At Blombos cave on South Africa's southern Cape, for example, archaeologists have recently discovered pieces of haematite (an iron oxide used to make red pigment) engraved with abstract designs that are at least 75,000 years old (Science, vol 323, p 569). Meanwhile, at the Skhul rock shelter in Israel, there are shell beads considered by some to be personal ornaments and evidence for symbolic behaviour as far back as 100,000 years ago (Science, vol 312, p 1785).

Further evidence may well come from caves elsewhere in the world, and indeed a tentative look at the existing records suggests that many of von Petzinger's symbols crop up in other places. The open angle symbol, for example, can be seen on the engravings at Blombos cave.

Does this suggest that these symbols travelled with prehistoric tribes as they migrated from Africa? Von Petzinger and Nowell think so. Davidson, on the other hand, who has identified 18 of these symbols in Australia, is unconvinced that they have a common origin, maintaining that the creative explosion occurred independently in different parts of the globe around 40,000 years ago. Instead, he thinks the symbols reveal something about a change in the way people thought and viewed their world, which may have emerged around this time. "I believe that there was a cognitive change, which suddenly put art into people's heads," he says.

Clottes, however, thinks they could be on to something. "Language and abstract thought were probably practised long before 35,000 years ago, since 'modern humans' are some 200,000 years old. We shouldn't be surprised by the sophistication of these people's thinking: they were our great-great-grandparents after all," he says.

But if people really did have a symbolic culture this far back, why don't we find more evidence pre-dating 40,000 years ago? "Perhaps the earlier symbols tended to be carved into perishable things such as wood and skins, which have now disintegrated," says von Petzinger. And even if they did paint in caves many of the rock surfaces will have eroded away by now.

Whenever these symbols did emerge, the acceptance of symbolic representation would have been a turning point for these cultures. For one thing, it would have been the first time they could permanently store information. "Symbols enabled people to share information beyond an individual lifespan. It was a watershed moment," says Nowell.

One huge question remains, of course: what did the symbols actually mean? With no Rosetta Stone to act as a key for translation, the best we can do is guess at their purpose. Clottes has a hunch that they were much more than everyday jottings, and could have had spiritual significance. "They may have been a way of relating to supernatural forces. Perhaps they had special symbols for special ceremonies, or they may have been associated with the telling of special myths," he says.

One intriguing aspect is their possible use in deception. "Once symbolic utterances are recognised, communication becomes more flexible," says Davidson. "One result is that ambiguity can be introduced for concealing truths."

With no key to interpret these symbols, though, we can't know whether ancient humans were giving false directions to rival tribes or simply bragging about their hunting prowess. Our ancestor's secrets remain safe - at least for now.

Doodler or da Vinci?

When our ancestors painted beautiful works of art, were they intending them to be viewed by others, or did they just paint for their own pleasure?

The Lascaux caves, in the Dordogne region of France, may have the answer. There you can see a painting of a red cow with a black head high on one of the walls. Up close the cow appears to be stretched from head to toe, but when viewed from the ground the cow regains normal proportions. This technique, known as anamorphosis, is highly advanced, and suggests the painter was considering his audience as he painted the cow.

Our ancestors probably took the quality of their work very seriously. Recent work by Suzanne Villeneuve, from the University of Victoria in British Columbia, Canada, shows that the images painted with the most skill tend to occur in places where large numbers of people would have been able to see them, while poorer-quality images were more likely to be in smaller cubby holes. In most cases it seems that only the "Leonardos" of the day were allowed to paint the big spaces.