The short answer: ritual killing probably did happen, but it can be proved far less often than popular accounts suggest. The famous figures come from Christian authors who were not there. The only detailed eyewitness was writing on the Volga. And the ground has yielded a small number of graves and wells that allow more than one reading. This article stays within the Viking Age, goes through the evidence text by text and find by find, and says what each method can and cannot do.
The subject here is roughly the 8th to 11th centuries, in Scandinavia and the regions Norse people reached. Bog bodies, weapon deposits and Roman reports on the earlier Germanic peoples belong to another period and are treated in Germanic sacrifice. For what an ordinary blót looked like, meaning a communal offering with a feast, and where it took place, see Blót sites. Normal grave customs are explained in Viking burial.
The question here is narrower: is there solid evidence that people of this period ritually killed other people? Texts by outsiders, bones in unusual positions and scientific analysis are available. None settles the matter alone.
The word sounds clear and is not. In the strict sense a sacrifice is a gift to a divine power. A woman killed so that she can follow her dead master fits that definition only partly, because she is first of all being sent with a dead person. Marianne Moen and Matthew J. Walsh acknowledge this in their 2021 paper “Agents of Death”. They still treat the killing of retainers as human sacrifice, understanding it as the offering of a human life to the dead person, to the gods, or as part of final rites.
Archaeology faces a harder limit as well. It sees a body and its position, not the reason for the death. Moen and Walsh stress that the cause of a deposition can never be known, only interpreted. A person with signs of fatal violence in someone else's grave may have been sacrificed. That person may equally have been executed, killed in battle, or have died elsewhere and been brought there later.
The best-known text stands in the fourth book of the history of the Hamburg church that Adam of Bremen wrote in the 1070s. He describes a temple at Uppsala and a festival to which the Swedes gather every nine years.
The sacrifice is of this kind: of every living thing that is male, nine head are offered […]. The bodies are hung in the grove that lies nearest the temple. […] Dogs and horses also hang there beside humans; their bodies, hanging mixed together, a certain Christian told me he had seen, 72 in number.Adam of Bremen, Gesta Hammaburgensis IV.27; my own translation from the Latin sentences given in Hultgård 2022, omissions marked
Three things stand out. First, Adam does not claim to have seen this himself; he relies on an unnamed Christian. Anders Hultgård points out that citing an eyewitness was also a literary device. Second, Adam breaks off on his own account: the chants at this rite, he says, are many and indecent and better left unmentioned. Third, 72 is a multiple of nine, a point that returns in the marginal notes.
These marginal notes, the scholia, are additions to Adam's text and are regarded as even less reliable. Scholion 141 says that the festival lasts nine days and that on each day one human is offered with other animals, making 72 creatures in all. Scholion 138 mentions a spring into which a living person is plunged. Scholion 140 reports that the Christian king Anund refused the prescribed sacrifice. Hultgård considers this last item credible, while he reads the spring episode rather as hearsay recast in terms of the medieval cold-water ordeal, a form of trial by God's judgement.
Scholars today read the account neither as reportage nor as pure invention. Rudolf Simek stresses that Adam's picture was already out of date when he wrote it, and argues that Adam drew on Thietmar's description of Lejre. Hultgård thinks a mixture is likely: genuine information, possibly from local oral tradition, combined with Christian polemical commonplaces. Among the latter he counts human sacrifice itself, the motif of appeasing the gods, and the restriction to male victims, which recalls biblical rules.
Specialists differ on the number nine. For Hultgård it is not mere stylisation: he treats it as genuine and points to the 7th-century Stentoften rune stone, which names nine he-goats and nine stallions, and to the nine nights of Hávamál. That stone, however, speaks of animals. A cultic nine does not yet prove human sacrifice in nines.
A point of language is also disputed. The Latin capita literally means “heads”, but it can simply mean “head” in the sense of a count of livestock. Hultgård considers decapitated victims in the grove very unlikely and prefers “items”, meaning whole victims. The often repeated version with nine severed heads therefore depends on a translator's choice.
The ground offers little help. According to Hultgård's survey, neither spring nor grove nor sacrificial remains have so far been found at Gamla Uppsala. The large Vendel-period hall had no comparable successor in the Viking Age, and the older idea of a cult building under the medieval church has not been confirmed. Absence does not prove Adam wrong, since a grove leaves few traces. It does not confirm him either.
About two generations before Adam, Bishop Thietmar of Merseburg wrote in the first book of his chronicle about a place called Lederun, today's Lejre on Zealand. There, every nine years in January, after the feast of the Lord's Epiphany, everyone is said to have gathered and to offer to their gods “LXXXX et VIIII homines”, 99 people, along with as many horses, plus dogs and cocks, the cocks standing in for hawks.
Thietmar presents this as something he has heard and speaks of old sacrifices. So he is looking back, not reporting as a witness.
Hultgård calls the number of victims exaggerated, to say the least. Ninety-nine people and 99 horses every nine years would also lead one to expect traces on a scale that I am not aware of from the excavations at Lejre; scholarship knows the place mainly as a hall site. Simek points to the lack of archaeological evidence, thinks a major sacrificial site at Lejre in the late 10th century most improbable and declines to trust Thietmar's account. Lejre was a centre of power, but for the number there is no independent evidence.
How the two reports relate remains unresolved. If Simek is right and Adam depends on Thietmar, then Lejre and Uppsala are no longer two independent witnesses to one pattern. If Hultgård is right and both reflect a genuine Norse order of nines, they support each other at least on the festival rhythm, though not on the numbers of victims. This cannot be clearly established.
The case of Ibn Fadlan is quite different. In 921/922 he travelled with an embassy from the caliph in Baghdad to the Volga Bulghars, and he describes the funeral of a leader of the Rus. The ritual ends with the burning of a ship on land. According to his account an enslaved woman declares herself willing to die with the dead man, is drawn into the proceedings over the following days and is finally killed. An older woman directs the rites; Ibn Fadlan calls her the “Angel of Death”. The text also has two men taking part. Our article on Ibn Fadlan gives the full sequence; what matters here is a killing the writer presents as part of the funeral.
The value of this source is high, because the author is present and, unlike Adam and Thietmar, has no church position to argue. Its limits are just as clear. He follows what people say only through an interpreter. He writes as an outsider. And the Rus on the Volga were a group whose origin and cultural mixture are disputed. They lived far from Scandinavia, and no one can say how far their practice carries over to Zealand or Lofoten.
Moen and Walsh warn against a further error. Modern readings of this single scene, they argue, have set what we expect of Viking Age sacrifice at all: a passive young woman, a powerful man. Their own counter-argument turns to the agents in the text, above all the “Angel of Death”, who alone sheds blood. The supposed willingness of the enslaved woman is also read in different ways: Moen and Walsh consider that it might be an attempt to take control of her own fate. Against that stand her legal position as an unfree person and the signs of sexual exploitation in the same report. None of these readings can be tested.
Multiple burials in which one body is furnished and the other is not, or shows signs of violence, are regarded as the main archaeological indication. They are rare at any rate: Moen and Walsh cite a compilation by Jensen (2016) with seven double burials of this type, one with a man and a female “victim”, two with a woman and a male “victim”, and four in which both are male. I cannot judge how complete that count is, but the pattern sits badly with the common picture in which mainly young women died with their masters.
From Flakstad in northern Norway come the remains of ten individuals; at least three of them lay without heads in double and triple graves. Elise Naumann studied the material in her dissertation and in a paper in the Journal of Archaeological Science. Carbon and nitrogen isotopes in bone collagen showed that the headless individuals had lived largely on fish, while others in the same graves ate much more land-based food and meat. Ancient DNA indicated that those buried together were most likely not related through the maternal line.
From this Naumann developed a hypothesis: the headless could have been slaves, decapitated before burial and given to their masters. She frames this explicitly as an assumption, and the limits of the data can be named. Isotopes measure diet, not legal status. Naumann found dietary differences even within one household, between men and women and between adults and children, and explains some of them by travel or fostering. Maternal DNA also does not exclude kinship through the father's line. The hypothesis remains plausible because several observations point the same way. It is not proven.
The grave at Gerdrup on Zealand held a woman with grave goods beside a man on whom Moen and Walsh, citing Bennike and Price, note evidence of hanging and bound feet. For a long time he was seen as an unfree man following his mistress. According to Moen and Walsh, David Wilson suggested the opposite: a man executed for murder or rape, buried beside his victim. Moen and Walsh do not regard that as untenable.
Then came the 2021 reassessment by Ole Thirup Kastholm and Ashot Margaryan. Both share the same mitochondrial DNA, the genome data point to a parent-child relationship, and the authors read the pair as probably mother and son. The data are thin because of very low coverage. How he died also remains circumstantial for them: the neck apparently broken, the ankles tied, though the body could have shifted after death. As an unfree servant of his mistress the man is now hard to maintain. A sacrificial reading is not disposed of, however: the authors lean towards the son as a companion killed for the burial, and also name other explanations.
Stengade on Langeland, grave FII, contains two men: one lying on his back and above him one lying face down with signs of violent death. According to Moen and Walsh, following Gardeła (2017), it is often presented as that of two executed criminals. The abstract of a paper by Jiří Starý (2005), by contrast, gives master and slave as the usual reading, describes one man as beheaded with bound legs, and also reports Zdeněk Smetánka's reading as fear of a revenant. So it is no clear case. Grave 55 at Lejre, also two men, is by contrast often read as a slave with his master.
The much-cited woman in the mound at Ballateare on the Isle of Man, with her large skull injury, is also less certain than is often said. According to Moen and Walsh it cannot be taken as secure that she was buried at the same time as the man in the main grave; Weale (2019) likens the injury rather to battle wounds, and Griffiths (2010) also allows for death by disease. For the Oseberg ship burial with its two women, sacrifice is repeatedly suspected but has not been demonstrated. According to a press report the anatomist Per Holck saw no reason to assume violence as the cause of death; the younger woman's broken collarbone had been healing for several weeks.
On the site of the Trelleborg ring fortress near Slagelse there are shafts that Poul Nørlund, in his 1948 publication, interpreted as a sacrificial site from before the fortress. A team led by Anne Birgitte Gotfredsen, with Primeau, Frei and Jørgensen, has re-examined three of them (Danish Journal of Archaeology). They date the shafts to the 8th/9th to 10th centuries. Shafts 47 and 121, with the children, predate the fortress built in 980/981; shaft 50, according to Nørlund, may still have been in use during the fortress period.
Shaft 47 held, near the bottom, the remains of two children of about four and seven years; shaft 121 the remains of two children of about four, plus fragments of an adult; shaft 50 skull and arm bones of an adult. Sex could not be determined. There are also many animal bones, including a young he-goat, a large dog, cattle and horse, along with objects. According to the authors these are the only known Danish Viking Age wells with the remains of children.
The strontium values in the tooth enamel of the four children fall within the Danish range, so the authors consider them most probably local. That speaks against the obvious guess that children were carried off from far away and deposited here.
The authors read the shafts as sacrificial wells and the children as deliberately deposited while still fleshed; the National Museum of Denmark speaks of human sacrifice without qualification. At the same time they state plainly that how the children died cannot be determined. Cut marks and trauma are absent, though that does not exclude violence. The bones were recovered by shovel without sieving, their positions were not systematically recorded, and only 48 percent of the originally identified animal bones could be found again. Assigning the goat to Thor is interpretation. So Trelleborg very probably shows ritual deposits of children. Whether they were killed for that purpose or were already dead, it does not show.
At Repton in Derbyshire the Great Army wintered in 873. In the 1970s and 1980s Martin Biddle and Birthe Kjølbye-Biddle excavated there, among other things, a converted Anglo-Saxon building holding the commingled remains of at least 264 people. For this subject a find outside that mound counts: a grave of four juveniles, a sheep's jaw at their feet, with a stone setting beside it that probably held a marker. The Antiquity paper gives their ages as 6 to 17, the University of Bristol press release as 8 to 18. At least two show signs of injury according to the release. The excavators read it as a ritual grave; Jarman and colleagues say it could have been a sacrificial deposit marking the completion of the mound, but that this is uncertain.
New radiocarbon dates by Cat Jarman and colleagues (Antiquity 2018) place this grave at 872 to 885, the time of the army. The date supports the connection, not the motive. Jarman herself thinks it very likely that the dead of the mound belonged to the army and stresses that this is not proven.
Scientific methods have improved the debate, but they do not measure what one is asking about. Strontium in tooth enamel gives clues to the area where someone grew up. Carbon and nitrogen in bone show whether a person ate mainly marine or land-based food. Ancient DNA shows kinship.
None of these methods recognises a sacrifice. They can weaken an interpretation, such as that of outsiders at Trelleborg or of mistress and servant at Gerdrup, and they can support one, as the dietary differences do at Flakstad.
Taken together, the evidence falls into grades. It is very well attested that people in the Viking Age killed animals ritually. Well witnessed, though only once and far away, is the killing of an unfree woman at a funeral in a milieu with Scandinavian roots. Probable, but not provable in any single case, are similar killings in some Norse graves. Possible and seriously discussed are ritually motivated killings of children at Trelleborg. Unconfirmed are the large, regular mass sacrifices described by Adam and Thietmar.
The finds lie in Lofoten, on Zealand and Langeland, in the English Midlands and on the Isle of Man, with a text from the Volga on top. They are spread over centuries and do not point to one shared rule. The background, by contrast, is clear: a society in which slavery was everyday and some people counted as property. That explains why researchers so often think of unfree people when they see accompanied burials. It does not replace evidence for the individual grave.
So did the Vikings sacrifice humans? Probably in individual cases, above all around funerals. Ibn Fadlan's account and some graves point that way. Regular mass sacrifices are attested only in texts that cannot be checked.
Is Adam of Bremen a reliable source? For Uppsala only to a limited extent. He depended on informants, had church interests, and according to Simek wrote when his picture was already out of date. Hultgård still considers some of his information genuine.
Why does the number nine come up so often? It appears in Adam, in Thietmar, on the Stentoften stone and in Hávamál. That suggests a role in Norse cult. It says nothing about how many people were killed.
Were the children at Trelleborg sacrificed? The authors of the study read the shafts as sacrificial wells and the deposit as deliberate. They also write that the manner of death cannot be established and that visible signs of violence are absent.
Was there human sacrifice in the Oseberg burial? That is debated because two women lay there together. There is no proof, and the younger woman's broken collarbone was already healing.
Germanic Sacrifice & Sanctuaries · Blot Sites: Horgr, Hof & Cult Hall · Viking burials · Ibn Fadlan & the Rus · Adam of Bremen.
Two spells on the blank leaf of a missal, entered in Old High German in the tenth century, discovered in 1841 by Georg …
Read more →Since 1961 Minnesota has fielded a football team with a Viking name, a horn on its helmet and a rallying cry that began …
Read more →The market is full of "Norse" symbols: on pendants, T-shirts, tattoos and wall art. Many are sold as ancient and …
Read more →A chieftain who talks about raids like quarterly targets, a slave who would rather stay a slave, and a Roman actor …
Read more →