A Blog by

Can You Tell a Woman by Her Handprint?

Edit, 12/14, 10:59pm: This post has now been updated with responses from the new study’s lead author.

A few months ago I wrote a story for National Geographic News that seemed to pique a lot of readers’ imaginations, and understandably so. It was about a study by Dean Snow reporting that, contrary to decades of archaeological dogma, many of the first artists were women.

Neat, right? But now there’s a twist in the tale: Another group of researchers is claiming the study’s methods were unsound. Snow has his own critiques of the criticism (more on that later). I’m less interested in who’s right than a fundamental question behind the controversy, and one that is relevant to all archaeological investigations: What does the present have to do with the past?

Snow’s study, published in the journal American Antiquity last October, focused on the famous 12,000- to 40,000-year-old handprints found on cave walls in France and Spain. Because these hands generally appear near pictures of bison and other big game, scholars had long believed that the art was made by male hunters. Snow tested that notion by comparing the relative lengths of fingers in the handprints. Why? Because among modern people, women tend to have ring and index fingers of about the same length, whereas men’s ring fingers tend to be longer than their index fingers.

Snow first scanned the hands of 111 people of European descent who lived near Pennsylvania State University, where he is an emeritus professor of anthropology. By comparing male and female hands on specific measures — such as the length of the fingers, the length of the hand, the ratio of ring to index finger, and the ratio of index finger to little finger — Snow developed an algorithm that could predict the sex of a given handprint. He also validated the algorithm on a second set of modern hands (50 males and 50 females).

The algorithm was only weakly predictive — with an accuracy of just 60 percent — because there’s a lot of overlap between the hands of modern men and women. But the equations were far more accurate when used on a set of 32 ancient hand stencils. The various measurements of these hands fell at the extreme ends of the modern sample, making it easy for the algorithm to categorize them as male or female. Snow found that 24 of the 32 prints — 75 percent — were female.

These hand stencils found in the El Castillo cave in Cantabria, Spain, were made by a man (left) and a woman (right), according to Snow’s study. Photos by Roberto Ontanon Peredo.

The new study, published Monday in the Journal of Archaeological Science, challenges Snow’s reference sample. A team led by Patrik Galeta of the University of West Bohemia in Pilsen, Czech Republic, collected handprints from 100 contemporary people in southern France and then ran those measurements through Snow’s algorithm.

Galeta found that Snow’s algorithm predicted female hands fairly well, but was useless for males, making it overall a bad predictor of sex. The study showed, in other words, that sex differences in hands among modern people living in Pennsylvania are not the same as differences among modern people living in France. “Our understanding is that hands of French males are on average smaller than U.S. males,” Galeta notes. And that, he adds, “is why U.S. methods failed to correctly identify French males.”

The bottom line: if two modern populations don’t match, then how can we possibly say anything about handprints tens of thousands of years old?

“What this shows is that a basic assumption that everyone has been making is wrong, which is that we can take a contemporary human population and use it as a model across space and time,” says archaeologist David Whitley of ASM Affiliates, an archaeological consulting firm in Tehachapi, California. Whitley was not involved in either study.

This might explain, Whitley adds, why researchers studying these old handprints have often come to contradictory conclusions. Before Snow’s work, evolutionary biologist R. Dale Guthrie performed a similar analysis of the cave prints and reported that most of them came from adolescent boys.

Snow, however, doesn’t agree with the criticisms of the new study. “I would stand by my guns here,” he says.

He sees two possible reasons that his algorithm didn’t work on the new French sample. One is that the Czech researchers didn’t use his algorithm in the same way that he did. Snow did his analysis in two steps, running the data first through an equation related to the length of the hand, and then running those results through another equation based on the ratio between the index and ring finger. The Czech researchers, in contrast, looked at the two equations separately.

Alternatively, it could be that the Czech researchers didn’t measure hand length the same way Snow did, he says. Snow measured from the tip of the middle finger to the creases where the wrist meets the palm. “If you measured the length of the hand using some other terminus at the base, you might lose a centimeter or so of the overall length,” Snow says.

So who’s right, and how can this be resolved? “I would have to see their data, and they would have to see my data, and we would have to work it out,” Snow says.

So far neither group has made contact with the other, though both parties seem willing. and the Czech group has not yet responded to my queries about their work. (If and when they do I’ll be sure to update this post.) The Czech group, for the record, rejects both of the explanations Snow proposed, saying that they used the algorithm and measured the hands exactly as Snow did.

Even if the Czech group is right, Snow says the main conclusion doesn’t change. “Even with their sample, they can show as well as I can that there were some women in them caves,” he says. “They might argue, well was it 50-50 or 70-30 or 80-20, but that part of it doesn’t concern me so much.”

Experts have been arguing over the identity of these handprints for decades, and that debate isn’t going away anytime soon. That’s part of good science. But I think this story also says something interesting about archaeology.

Archaeologists are constantly turning up objects from the distant past, and their job is to figure out what (or, in this case, who) they were. They begin, naturally, by making assumptions based on the objects and people we’re familiar with today. “It’s an issue we always confront — making ‘presentist’ projections onto the past,” Whitley says.

In the case of these handprints, the projection relates to our bodies. But it could be anything. “If you find a pot, then just calling it a pot assumes you have some understanding of what it was,” Snow says. “We all make inferences. You just have to be reasonably comfortable with your inferences.”

6 thoughts on “Can You Tell a Woman by Her Handprint?

  1. All “science” that reaches conclusions about past items and events which were not observed and recorded by humans is prone to error, and at best we can only estimate how wrong we might be. In this case, we have what amounts to human records, but in a crude form that leaves much to interpretation. The original assumption that the hand prints were all male wasn’t even based on present behavior, just male-centric thinking. Even the hunting scenes may have been painted by women, either because they attended the hunts (perhaps hanging back during the kills, as, from what I recall, their skeletons don’t show as much damage as the males), or because they were out and about enough to know what the animals looked like and depict what their men described.

  2. All “science” …. is prone to error, and at best we can only estimate how wrong we might be.

    Fixed that for you, David. Thing is, we have a whole branch of mathematics aimed specifically at dealing with such estimates.

    1. Good points, but most areas of science involve things we’ve been able to observe directly many times and demonstrate by various tests and finally turn into reliable applications, technology, etc. In some cases, many lives are depending on these things to be within very fine error limits.

      In other areas of science (not many, not so large, and generally not so relevant), our estimates and even mathematical analyses depend on making assumptions that we can’t really be sure of, and there’s no way to make repeated observations or such a wide variety of tests. It may be more of a spectrum than black or white, true, but let’s not deny there are differences.

  3. All science is about things we’re able to observe directly. Yes, the events that left Sue the T. rex entombed in rock are long past. But everything we know about her comes from what people observed when they dug her out of that rock. Events in the past leave traces in the here and now, and that’s what we’re doing science on.

    I would argue that much of modern materials science, physics, etc. operates the way you’re describing at least as much as historical sciences such as paleontology or astronomy. After all, when they found the Higgs boson, no one saw the darn thing – they didn’t even see traces in a bubble chamber. What they saw was a little bump in a graph drawn from data taken from a bunch of detectors in the middle of a giant ring of magnets. In order to accept that this means a Higgs boson, we need to accept that the software works right, that the detectors actually detect what we say they do, that all of the connections work right, that those actually are some mysterious things called protons that are actually being smashed together in the middle of those magnets, that someone actually did the calculations right to figure out what signal those bosons “should” show……..

    Our entire view of the world is built upon theoretical understandings that could potentially change tomorrow.

    And the idea that things like archaeology, paleontology, or evolutionary biology aren’t relevant is highly problematic, to my mind. Oil geologists use paleontology to date the rocks they’re looking for oil in — sounds relevant to me. Those few people out there looking for new antibiotics are responding directly to evolution of bacteria — pretty darn important stuff. (The many more people researching new pesticides or alternatives are doing the same thing, of course.)

    1. Again, I quite agree there are other areas of science where we need to be cautious not to be building a house of cards or “six feet out on the fog.” Likewise, there are a number of things about the past that the traces left behind clearly reveal. I don’t mean to smear all of any field, archaeology, paleontology, or evolutionary biology — they all have aspects which involve things that can be observed, We can see clothing in cave drawings comparable to those worn by Otzi the Iceman, for example (http://www.ancient-wisdom.co.uk/caveart.htm), and the depths into the caves and the pigments used are solid facts. Likewise many other artifacts that we don’t have to “estimate” were made by humans.
      As to relevancy, again, there’s varying relevancy in different areas of each field. I don’t think the examples you’ve given are all that strong, though. Oil geologists don’t really need to “date” the rocks, they simply look for a pattern that shows the proper sequence of types of rock, especially domes and impermeable and permeable formations, right? Also, evolutionary biology didn’t predict that we would have to start dealing with resistant microbes (or pests, or weeds), and on the other side of the coin, if evolutionism hadn’t become popular, people would have just spoken of dealing with new breeds of bacteria or whatever, comparing them to new breeds of dogs, pigeons, etc.

  4. There is no mention of the hand size of actual human bones of earlier people. I know that anthropologists can determine if skeletons are male or female. I would think that the relationship of hand size/shape and gender of these earlier people would be more applicable to the people who created the cave art.

Leave a Reply

Your email address will not be published. Required fields are marked *