Birka's warrior was female in the 1970s. Nobody accepted it until 2017.

In 1878 the Swedish archaeologist Hjalmar Stolpe dug up a chamber grave on the island of Bjorko, at the Viking-age trading town of Birka. He catalogued it as Bj 581, and it is one of the richest warrior burials ever found in Scandinavia.

The list of what was in it is worth reading slowly. A sword. An axe. A fighting knife. Two spears. Two shields, propped against the walls at the head and the foot. A quiver of twenty five armour-piercing arrows. Two horses, one bridled. And, resting in the lap of the body, a bag holding twenty eight gaming pieces, three antler dice and a king piece marked with an iron nail, with an iron-framed gaming board propped up alongside.

The grave sat on a raised terrace directly beside the garrison and the hillfort. For a hundred and forty years, nobody had the slightest doubt about what it was: a high-ranking professional warrior, probably a commander, given the gaming set. As the osteologist Anna Kjellstrom later put it, "Only a few warriors are buried with gaming pieces, and they signal strategic thinking."

There was one detail nobody had checked carefully. The person in the grave was a woman.

Forty years of not hearing it

This is the part that makes Bj 581 more than a nice piece of trivia about Viking women.

An osteological analysis first identified the skeleton as female in the 1970s. The finding was dismissed.

In 2016, Anna Kjellstrom of Stockholm University examined the bones again and reached the same conclusion again: female. Pelvis, mandible, the standard skeletal markers.

In 2017, a team led by Charlotte Hedenstierna-Jonson of Uppsala University published a genomic study in the American Journal of Physical Anthropology. No Y chromosome, two X chromosomes. The matter was closed.

So the bones had been saying the same thing for roughly four decades, and it took DNA before the field would hear it. That gap is the story.

The objection that gave the game away

When the 2017 paper landed, the response was not "how interesting". It was a serious scholarly argument that the interpretation of the grave could no longer stand: that grave goods do not prove occupation, that weapons in a burial might be inherited wealth or family display rather than a record of what the dead person did.

That objection is not stupid. It is, in the abstract, completely correct. The problem is when it arrived.

Since it was excavated in the 1870s, it has constantly been interpreted as a warrior grave because it looks like a warrior grave and it's placed by the garrison and by the hillfort. Nobody's ever contested it until the skeleton proved to be female, and then it was not a valid interpretation anymore.

That is Hedenstierna-Jonson, and it is the cleanest statement of the whole problem I have come across. Not one piece of evidence about the grave changed between 2016 and 2017. The sword did not change. The arrows did not change. The gaming set did not change. The only thing that changed was the sex of the body, and at that exact moment the standard of proof went up.

The fallback position was that the bones must be the wrong bones, mixed up in storage over a century of museum handling. It is a reasonable worry, and it fails on a small, wonderful detail: Stolpe was a fastidious excavator, and at the moment of digging he had written "Bj 581" onto every single bone in India ink.

To be fair to the critics, some of the scepticism predates all of this and is not about sex at all. Anne-Sofie Graslund argued back in 1980 that calling any Birka grave a "warrior grave" overreads it, since a pile of weapons may mark a social elite rather than a working soldier. Judith Jesch's criticism of the 2017 paper focused on its use of written sources and its failure to weigh alternative readings. The authors published a full response in Antiquity in 2019 and did not budge.

Those are real arguments. They are just arguments that, for a hundred and forty years, nobody thought worth making.

The princess who was reassigned as a priest

If Birka is the best documented case, the grave at Vix in Burgundy is the most extraordinary.

Excavated in 1953 by Rene Joffroy, it held a Celtic burial of staggering wealth: a solid gold torc, a four-wheeled wagon, and an imported Greek bronze krater more than a metre and a half tall, the largest metal vessel to survive from the ancient world. The body was identified as female at the time.

What happened next is close to unbelievable. In a 1983 book, the archaeologist Konrad Spindler resolved the difficulty of a woman holding that much power by proposing that the occupant was not a woman exercising authority but a man in women's clothing: a transvestite priest.

Read that again. Faced with a female body and elite grave goods, the explanation reached for was a cross-dressing male religious official. A third gender category was easier to accept than a female ruler.

The archaeologist Bettina Arnold took this apart at a 1989 conference and in a paper with a title that does the work on its own, "The Deposed Princess of Vix". Her point was not that Spindler was uniquely foolish. It was that the discipline had a habit of writing powerful women out of the past, and that the habit was invisible to the people practising it, because it felt like caution.

Why this kept happening

Underneath both cases sits a circular method that was, for a long time, standard practice.

Skeletons are hard to sex, especially fragmentary ones. Grave goods are easy to read. So excavators fell into the habit of sexing the body by what was buried with it. Weapons meant a man. Jewellery and spindle whorls meant a woman. The rule was applied, the results were tabulated, and the tables then proved that men were buried with weapons and women with jewellery.

The circularity is total. The conclusion was assumed at the point of measurement, and then recovered from the data as though it were a discovery. Nothing about that process could ever have produced a female warrior, no matter how many were in the ground.

Peru, and a genuinely new kind of evidence

In 2018, at a site called Wilamaya Patjxa high on the Andean altiplano in Peru, excavators found a burial from between 9,000 and 8,700 years ago. The individual, a young adult, had been interred with a big-game hunting toolkit: projectile points and the tools for butchering and processing hides. Aymara community members later gave her a name, Warawara, meaning star.

Proteomic analysis of tooth enamel established she was female. That method matters, because it does not depend on eyeballing a pelvis or on what was lying next to her.

Randall Haas and colleagues published the find in Science Advances in 2020 and then went looking for company. They combed published records of late Pleistocene and early Holocene burials across the Americas, 429 individuals from 107 sites, and found 27 buried with big-game hunting tools. Eleven of those were female, fifteen male. From that they estimated that somewhere between 30 and 50 per cent of big-game hunters in these populations were women.

And now the part that complicates it

Here I have to be even-handed, because the corrective has its own problem, and it is the same problem.

Robert Kelly of the University of Wyoming, who has no quarrel with the Peruvian burial itself, pointed out the obvious difficulty with the wider survey: tools in a grave do not prove the occupant used them in life. They can be offerings. They can be gifts from the living. And in Haas's sample of female "hunters", two were infants.

An infant buried with a spear point is not evidence of an infant hunter. It is evidence that somebody put a spear point in a baby's grave.

Notice what has happened. The old error was reading identity off grave goods: weapons, therefore male. Part of the new reading does the same thing in the opposite direction: weapons, therefore hunter. It is the same inference, and it is no more sound for arriving at a conclusion we prefer.

The pattern repeated in the ethnographic literature. A 2023 paper in PLOS ONE titled "The Myth of Man the Hunter" reported that women hunted in 79 per cent of the foraging societies it examined, fifty out of sixty three. It travelled around the world in a week.

In 2024, Vivek Venkataraman and colleagues went through the data. They found the analysis was not reproducible, that the sampling criteria were unclear, that 35 per cent of the societies in the sample had not come from the database the authors said they had used, and that the coding contained numerous errors. Re-coding the same societies, they found that women rarely or never hunted in about a third of the cases the original paper had counted as women hunting. PLOS ONE published a correction.

Their own conclusion is the honest one, and it is in their title: female foragers sometimes hunt, and gendered divisions of labour are still real.

What the record actually supports

So, to answer the question plainly. There is no tally of times archaeologists refused to accept a female skeleton, and anybody who gives you a number is making it up. What there is instead is a documented, named, dated pattern, with at least two cases where the refusal is beyond argument. A woman in Sweden was identified by her bones in the 1970s and had to wait forty years and a genome for anyone to act on it. A woman in France was reassigned as a man in a dress rather than be allowed to have been in charge.

What that record supports is not the claim that women were secretly doing all the hunting and fighting. It supports something narrower and more damning: that for a long time the discipline let its assumptions do the work of its evidence, and did not notice, because the answers kept coming out the way everyone expected.

The useful lesson is not about Vikings. It is that the moment a piece of evidence starts producing an answer you did not expect, you find out very quickly whether your standard of proof was ever about the evidence at all. Birka failed that test for forty years. The temptation now is to fail it in the other direction, and call that progress.