The question is asked constantly and answered badly in both directions. The evidence is better than the popular arguments suggest — and messier.
Published in partnership with Dude-Hack.
Icelandic law was explicit on the point. Among the provisions of Grágás is one confirming that a man was <cite index=”183-1″>entitled to purchase a slave specifically for the purpose of sexual pleasure</cite>.
That is not an inference drawn from a saga episode or an outsider’s travel diary. It is a legal provision, and it is the hardest single piece of evidence bearing on a question that surfaces in every discussion of Norse thraldom: what happened to enslaved women, and was sexual coercion part of it.
The honest answer requires sorting four unequal bodies of evidence, each strong where the others are weak, and each carrying problems that popular retellings tend to skip.
The law codes
The Grágás provision above appears in a legal culture that took sexual offences seriously — but in a particular way. Where a free woman was concerned, offences from kissing upward carried penalties running to outlawry. Crucially, these were treated as wrongs against her male kinsmen rather than against her. It was they who prosecuted, and to them that compensation was paid.
An enslaved woman had no such kinsmen, and that absence is the whole architecture of her position. The protective mechanism ran through family, and she had been removed from one.
Legal status was not fixed even so. Several sources, including medieval law codes, suggest a concubine might hold the same status as a slave — while other scholarship argues that many grades of concubinage existed, some carrying real influence within networks of political obligation.
The obvious objection to using these codes is that they were written down after the conversion, centuries after the practices they describe. The counter-argument, set out by Ben Raffield, Neil Price and Mark Collard in their 2017 study of polygyny and concubinage in Viking-Age Scandinavia, is straightforward: a Christian legal culture committed to monogamy had no reason to invent provisions accommodating concubinage. If something must be expressly forbidden or expressly regulated, people were doing it.
The sagas
The sagas of Icelanders supply the individual cases, and the most cited is Melkorka in Laxdæla saga. An enslaved Irish woman, said to be a king’s daughter, she is bought by Hoskuldr as a concubine in chapter twelve. In Vatnsdæla saga a married man maintains a formalized relationship with a concubine who bears his child — and, in the same episode, his wife orders that child exposed.
That second detail is worth pausing on, because it shows the structure rather than the incident. A concubine’s child had no secure standing against the interests of a legal wife.
The source-critical problem here is severe. These texts were composed in Christian Iceland two or three centuries after the events they narrate, by authors with their own moral commitments. Raffield and his co-authors note that such episodes are strikingly infrequent, which may itself reflect the standards of the writers rather than the practices of their ancestors. What can be taken from the sagas is not literal reportage but consistent patterns of description — and what medieval Christians evidently found believable about the pre-Christian past.
The outsiders
The most direct testimony comes from travellers who had no stake in how Scandinavians were remembered.
Ahmad ibn Fadlan, encountering Rus traders on the Volga in the tenth century, describes enslaved women being sold primarily as sexual commodities: he singles out youth and beauty as the qualities that mattered to buyers, and describes their repeated exploitation by the dealers holding them. He records a Rus ruler with forty concubines, and a retinue of four hundred warriors each provided with two enslaved women.
Adam of Bremen, writing around 1070, described Swedish practice in similar terms — a man taking two or three wives or more according to his means, with rich men and princes taking an unlimited number. The skaldic poem Hrafnsmál has King Haraldr rewarding his warriors with gold and slave-women from the eastern lands, which places the practice inside the Norse world’s own account of itself rather than only in foreign observation.
Two caveats apply. Whether the Rus can be treated straightforwardly as Scandinavians is contested — they were an ethnically mixed group, however substantial the Scandinavian component. And outside observers writing about distant peoples had their own purposes.
But the convergence is the point. Legal texts, sagas, Arab travellers and a German cleric, written in different centuries, languages and traditions, describe the same institution.
The objects, and the silence
Material evidence is where the picture thins almost to nothing.
The Swedish History Museum’s account of free and unfree status in the Viking Age is blunt about it: the proportion of thralls in the population cannot be estimated, and physical traces amount to <cite index=”184-1″>occasional finds of collars, shackles, or iron fetters</cite>. Neck irons and shackles from Birka and from Skedala in Halland are on display in Stockholm, and they are close to the whole corpus.
The museum makes a further point that deserves wider circulation: iron was valuable, and rope was not. People could be restrained perfectly well with material that leaves no archaeological trace at all. Absence of shackles is not absence of captives.
There is also genuine counter-evidence worth stating. Anne-Sofie Gräslund’s study of 959 runic inscriptions from Uppland found not one describing a man with two or more wives simultaneously. That looks decisive until one notes that those stones are overwhelmingly eleventh-century and Christian, raised in a context where the Church had made such arrangements unmentionable, and that a polygynous household would in any case name only its senior wife in an inscription concerned with inheritance.
The strand that is not a text
One line of evidence passes through no medieval author at all.
When deCODE Genetics sequenced 25 ancient Icelandic skulls from the first generations of settlement and compared them with modern populations, the founders proved to be a near-even mixture of Norse and Gaelic ancestry. Norse ancestry then rose from 57% at settlement to roughly 70% today.
The researchers’ own suggested explanation for that drift is the lower reproductive success of the Gaelic settlers, <cite index=”185-1″>many of whom were taken to Iceland as slaves</cite>. Later immigration from Denmark may also have contributed.
An earlier generation of genetic work had pointed the same way but relied on modern Icelanders rather than Viking-Age remains — a limitation the 2018 study removed.
So what is the answer?
Sexual access to enslaved women was legally sanctioned, socially normal, and attested across every independent category of source available. On the substance of the question, there is no serious scholarly dispute.
The phrase “sex slave”, though, flattens something the evidence does not support. Thraldom was not a single condition. The same system that produced the trade Ibn Fadlan watched also produced concubines with standing in political alliances, freed thralls who raised runestones, and children of enslaved mothers who inherited from free fathers. Most captives were not kept at all — Scandinavians were traders in people more than holders of them, and the goal was silver.
And a woman in that system could occupy several positions at once: property in law, mother to a free man’s heir, and a person whose descendants would be indistinguishable from anyone else’s within three generations.
The texts have been read about as far as they can be. What will move the question next is isotopic and genetic work on the burials themselves — which is, appropriately, the only source that was never written by someone with a reason to shape the story.
