Saga Period Iceland


Iceland is known to men as a land of volcanoes, geysers and glaciers. But it ought to be no less interesting to the student of history as the birthplace of a brilliant literature in poetry and prose, and as the home of a people who have maintained for many centuries a high level of intellectual cultivation. It is an almost unique instance of a community whose culture and creative power flourished independently of any favouring material conditions. and indeed under conditions in the highest degree unfavourable. Nor ought it to be less interesting to the student of politics and laws as having produced a Constitution unlike any other whereof records remain and a body of law so elaborate and complex, that it is hard to believe that it existed among men whose chief occupation was to kill one another.

            ( James Bryce, Studies in History and Jurisprudence 263 (1901))


About forty years ago, two prominent economists at the university of Chicago published an article that pointed out a problem with the conventional system for enforcing criminal law.[1] I like to summarize their argument with a brief story.

I am a cop, you are a criminal. I have the goods on you, the evidence that will send you to jail. The cost of being convicted and jailed is, to you, the equivalent of a hundred thousand dollar fine. The benefit to me of getting you convicted is a gold star on my report card, a boost to my career that will raise my lifetime income by ten thousand dollars. Seen from the perspective of Dragnet, the outcome is obvious: I turn over the evidence to the D.A., you go to jail. Seen from the perspective of economics, it is also obvious. I have something worth ten thousand dollars to me and a hundred thousand to you. Markets move assets to their highest valued use. You pay me something between ten and a hundred thousand dollars and I burn the evidence.

To keep this from happening, to make the system work as designed, we will need a second layer of cops watching the first layer—and perhaps a third layer watching them. That raises the cost and complication of the system. Becker and Stigler suggested a simple solution—replace the salary of the policeman with a reward. When you pay a hundred thousand dollar fine or receive the equivalent prison sentence, I get a hundred thousand dollars. Now the only bribe I am willing to accept is for at least a hundred thousand dollars, which imposes the proper punishment on you while saving the cost of a trial.

Two prominent legal scholars, also at the University of Chicago, responded, pointing out issues that the first set of authors had failed to deal with, including the question of who had the right to catch a criminal and collect the reward.[2]  One possible solution was to make it a property right of the victim, a claim against the criminal. At which point, as Landes and Posner pointed out, Becker and Stigler had reinvented the tort system. Under tort law, as under the hypothetical version of criminal law, the victim has a claim against the offender, collected by an in court conviction or out of court settlement.[3]

I got interested in the exchange and added two articles to it. One offered a solution to a technical problem with a system of privately prosecuted criminal law that Landes and Posner had pointed out.[4] The other described an actual legal system similar to the imaginary one they were discussing. In saga period Iceland a thousand years ago, if you killed someone his relatives sued you. That system was what first got me interested in the broader subject of this book. This chapter is a much revised version of that article.

I. The Problem of Sources

Our knowledge of the Icelandic legal system is based on sources of two sorts: the sagas, histories and historical novels written down in the thirteenth and fourteenth centuries, and Gragas, a collection of legal texts written down in the late thirteenth century. When I first tried to make sense of the system, Gragas had not yet been translated into English. Since I did not read Old Norse, I based my article on the sagas and the secondary literature.[5] Returning to the subject while writing this book, one of the first things I did was to read Gragas, now available in English. I discovered that parts of it were inconsistent both with my old account of the system and, as best I could tell, with the sagas.

To check the latter conclusion, I read though all of the sagas set in Iceland[6] and concluded that while my article had been mistaken about some legal details, I had for the most part correctly described the system as shown in the sagas. [7] Gragas was inconsistent not only with the family sagas, written down two or three centuries after the events they described, but also with the Sturlung sagas, whose authors were describing events many of which occurred during their lifetime, some of which they were participants in.[8]

The most important inconsistency had to do with out of court settlements. According to Gragas, an offence for which the legal penalty was full outlawry could be settled on less harsh terms only with the approval of the Lögrétta, the law council. Most killings in the sagas were resolved by settlement, but I found no examples of settlements either being approved by the Lögrétta or blocked by its failure to approve. Many settlements occurred when the Althing was not in session and so the Lögrétta, which met only at the Althing, did not exist to approve or disapprove them.[9]

According to Gragas, once someone had been charged with a serious offense it was illegal for anyone to harbor him—give him food or shelter. Equally strange, at least to modern sensibilities, it was illegal for him to attend any assembly, including the one at which he was to be tried.[10] His defense had to be conducted by someone else in his absence. I have found only one passage in the sagas where someone was penalized for harboring a killer before he had been tried and convicted; that was part of a settlement, not a court verdict.[11] Defendants often but not always appear at the assembly at which they are being tried.

Gragas specifies acts by which someone forfeited his immunity, could be killed without legal consequences. That included any attack, even a blow that failed to land, and applied not only to the perpetrator but to companions who knew of or assisted the intended attack.

It is prescribed that a man on whom injury is inflicted has the right to avenge himself if he wants to up to the time of the General Assembly at which he is required to bring a case for the injuries; and the same applies to everyone who has the right to avenge a killing. Those who have the right to avenge a killing are the principals in a killing case. The man who inflicted the injury falls with forfeit immunity at the hands of a principal and at the hands of any of his company, though it is also lawful for vengeance to be taken by other men within twenty-four hours.

The legal pattern in the sagas is more complicated. There are passages where an attacker is held to have fallen with forfeit immunity.[12] There is one that seems to imply that immunity is forfeit only if someone on the other side was killed or wounded.[13] There are others where the attack is treated as a separate tort, with damages to be set off against those owed for killing the attackers.[14] But in most cases where there is an unambiguous attack by one person or group against another the rule of forfeit immunity is ignored, with deaths on each side set off against deaths on the other side and wergeld or outlawry owed for any excess deaths.[15]

According to Gragas, almost anything of a sexual nature between people not married to each other was a serious offense.

“If a man kisses a woman in private, with no one else present and with her consent, then he incurs a penalty of three marks, and the case lies with the same man as an intercourse case would. But if she takes offence at it, then the case lies with her and the penalty is lesser outlawry. If a man gives another man's wife a secret kiss, the penalty for that is lesser outlawry whether she allows it or whether she forbids it, ... If a man asks a woman to sleep with him, the penalty for that is lesser outlawry.” (K § 155 G2 p. 69)

But Gragas also gives rules for inheritance which take for granted the existence of illegitimate children. In some cases the mothers might have been vagrants, to whom the same rules did not apply, or thralls, [16] or the father might have been outlawed. But one of the bishops of Iceland was illegitimate and his mother was a sister of the previous bishop, clearly not a thrall or a vagrant.[17] In the Sturlung sagas practically every important man has a mistress; one chieftain is in bed between his two mistresses when attackers show up.[18] The mistresses are described as the daughters of respectable farmers.[19] Being the mistress of an important man was viewed by some as better than being the wife of someone less important. There are cases in the sagas where someone is charged with an intercourse offense, but with only a few exceptions they are cases where intercourse led to pregnancy.[20] 

What explains conflicts between Gragas and the sagas? I can see at least four alternatives:


1. The sagas are wrong. They do not accurately describe how the legal system functioned in the period they cover.

2. Gragas is wrong.

3. Gragas describes the system as it existed at the end of the period, the sagas as it existed in the first century and a half.

4. Gragas describes the legal system as it existed on paper but not as it existed in practice.


The first alternative gets us into an old scholarly controversy. The family sagas describe events in the tenth and early eleventh centuries but their written texts date from the thirteenth and fourteenth. One possibility is that they were composed shortly after the events they describe, passed down in oral form for two or three centuries, then committed to writing. An alternative is that they were composed by the people who wrote them down, based on bits and pieces of tradition. If the latter is correct, their picture of the legal institutions of the earlier period might be no more reliable than the picture of the wild west in modern westerns or of medieval Europe in bad historical novels.[21]

One argument against the oral tradition version is that the sagas are prose, not verse, hence more easily mutated in transmission.[22] On the other hand, the authors of many of them had lived under the legal system in which they were set, even if a later version—the major changes occurred after 1263, when the Icelanders agreed to turn over authority to the king of Norway.

Many years ago Jesse Byock, a leading American scholar of the sagas, published an ingenious piece of evidence in, of all places, Scientific American. He demonstrated that a collection of apparently unrelated details about Egil Skallagrimsson, the central figure in Egil Saga, and his father and grandfather all fit the theory that they suffered from Padgett’s Syndrome, a hereditary disease first identified in the 19th century. That makes sense if the saga was composed at a time when the relevant details were still in living memory, less sense if it was created two or three hundred years later.[23]

Byock’s view is that while details of the saga plot might be recreated each time they were told, the stories were based on a historical and institutional background known to both teller and audience. [24]  Eventually someone did it one more time in writing. That suggests that they should be reliable as a source of information on institutions if not always on historical details. Sigurdsson takes a similar position.[25]

Whether or not the family sagas can be taken as reliable historical documents, it is generally agreed that the Sturlung sagas, dealing with events much closer to their composition, can be. Hence where a rule in Gragas contradicts the Sturlung sagas, I think we can be reasonably confident that it does not describe the actual law code, at least in practice.

The second alternative gets us to the nature of Gragas. It was assembled from two long accounts of the law and several fragmentary ones, all presumably compiled by private individuals for their own use. The earlier account was written down about 1260, just before the end of the period of Icelandic independence, the later about 1280, after Iceland had come under Norwegian rule. One rule included is that, in case of conflict among written accounts of the law, the texts belonging to the bishops have priority, which implies that written accounts of the law differed.[26] So the Gragas texts may represent some mix of what the law was, what the person writing it thought it was and what the person for whom it was written wanted it to be.[27]

The third alternative, that the law changed over time, is true in at least one respect. Iceland went Christian in the year 1000. Gragas  contains a long section dealing with churches, what things you are forbidden to do on a holy day and the penalties for doing them, and related issues. But the comparison of the Sturlung sagas to the family sagas suggest that although law in practice changed as the system broke down, law in theory, aside from the Christian additions, remained very much the same.

The fourth alternative strikes me as a plausible explanation in the case of intercourse offenses. Even if seduction was an offense that the woman’s father could prosecute, it might have been more prudent not to to, especially if no pregnancy had occurred. In one case we are told that the seducer responded to the father’s complaint by offering to marry his mistress—and did so.[28] It is a possible, but somewhat less plausible, explanation of the conflict between the penalties imposed, according to Gragas, on a defendant prior to his trial and the behavior of defendants in the sagas. This is more of a problem for the family sagas than for the Sturlung sagas, since by the Sturlung period defiance of the law by powerful individuals had become common.

There is one additional inconsistency between Gragas and the sagas, in particular the Sturlung sagas, that supports the idea that Gragas  is in part a wish list. In the Sturlung sagas, parties to legal disputes quite often arrive at the Althing with a substanial army, hundreds of supporters—the larger the force, the better the odds of a favorable outcome. A common pattern is for neutral parties, sometimes led by one of the bishops, to intervene to find some acceptable compromise in order to prevent a battle.

According to Gragas, no litigant can bring more than ten men to court.[29]

I conclude that the claim in Gragas that settlement required the consent of the Lögrétta is inconsistent with massive evidence in both the family and the Sturlung sagas, hence almost certainly false. I take the rules on forfeit immunity and intercourse offenses as existing in some form but not consistently enforced. Where Gragas is not contradicted by events in the sagas, I take its legal rules as a reasonable guess at the rules in force.


In the latter half of the ninth century, King Harald Fairhair unified Norway under his rule. A substantial number of the inhabitants, unhappy with the change, left;[30] many went either directly to Iceland, which had been discovered by the Norse a few years earlier, or indirectly via Norse colonies in England, Ireland, Orkney, the Hebrides, and the Shetland Islands. The political system which they developed there was based on Norwegian[31] traditions  with one important innovation—there was no king. The relationship between the Icelandic goði and his thingmen (thingmenn) was contractual but not territorial. The goði had no claim to the thingman's land and the thingman was free to transfer his allegiance.[32]

At the base of the system stood the goði (pl. goðar) and the goðorð (pl. goðorð). The original goðar seem to have been local leaders who built pagan temples and served as their priests. A goði received temple dues and provided in exchange both religious and political services. The goðorð was his congregation.

Under the system of laws established in A.D. 930, these local leaders were combined into a national system. In 960, Iceland was divided into four quarters,  each quarter containing nine goðorð clustered in groups of three called things. In 965 three more goðar were added in the North Quarter and in 1005 an additional three “new goðar” in each of the southern, eastern and northern quarters. The new goðar had seats in the Lögrétta and played a role in appointing judges to the fifth court but not to the quarter courts.[33]

The one permanent official of this system was the lögsögumaður or lawspeaker; he was elected every three years by the inhabitants of one quarter, which quarter it was being chosen by lot. His job was to memorize the laws, to recite them  once during his term in office, to provide advice on difficult legal points and to preside over the Lögrétta, the legislature.

The members of the Lögrétta were the goðar, new and old, for each of these two advisors, plus, after Iceland went Christian in 1000, the two bishops. Decisions were made by majority vote subject to attempts to first achieve unanimity.

The laws passed by the Lögrétta were applied by a system of courts, also resting on the goðar. The thing court or Vorþing was held at the spring Assembly of each quarter. The judges[34] were chosen twelve each by the goðar of the thing, making thirty-six in all. Next came the quarter-thing for disputes between members of different things within the same quarter; these seem to have been little used and not much is known about them.[35] Above them were the four quarter courts of the Althing (alþingi) or national assembly, an annual meeting of all the goðar, each bringing with him at least one-ninth of his thingmen. There were procedures by which a party to a suit could veto its settlement in the Vorþing, forcing it up to the appropriate quarter court. Above the quarter court, after the reforms credited to Njal, was the fifth court. Cases undecided at any level of the system went to the next level; at every level the judges were appointed by the goðar, each quarter court and the fifth court having judges appointed by the goðar from all over Iceland.[36] The fifth court reached its decision by majority vote; the other courts seem to have required that there be at most six (out of thirty-six) dissenting votes in order for a verdict to be given.[37]

A court system requires some way of determining the facts of the case, made more difficult in a society where most people were illiterate hence written documents of limited use. The Icelanders made use of a system of panels of varying size. In some contexts the panel consisted of the nine nearest neighbors to the site of an event, such as a killing. In others, it consisted of witnesses to a required legal act. There were requirements for who could be on a panel and procedures for dismissing unqualified members, such as anyone too closely related to plaintiff or defendant.

The goðorð itself was two different things. It was a group of men—the particular men who had agreed to follow that goði, to be members of that goðorð. Any man could be challenged to name his goðorð and was required to do so, but he was free to choose any goði within his quarter and to change to a different goðorð at will. [38] The goðorð was also a bundle of rights, including the right to sit in the lögrétta and appoint judges for certain courts. Perhaps most important, it was the right to be the person through whom ordinary farmers plugged into the legal system. The farmer who was the thingman of a goði owed him an annual thingtax, used to pay the expenses of those thingmen who accompaned the goði to the Althing. The amount of the thingtax was negotiated between goði and thingman.[39] The goðorð in this second sense was marketable property. It could be given away, sold, held by a partnership, inherited. Seats in the law-making body were quite literally for sale.

I have described the legislative and judicial branches but have omitted the executive. So did the Icelanders. The function of the courts was to deliver verdicts on cases brought to them. That done, the court was finished. For serious offenses, conviction meant full outlawry. The outlaw’s property was confiscated, part going as a damage payment to the victim or his heirs, part to support the outlaw’s dependents. If more is available, the goði in charge of the confiscation court gets a a cow or a four year old ox. The remainder is divided between the prosecutor and the men of the quarter if he was outlawed at the Althing, the men of the assembly for the district where the confiscation court was held if he was outlawed at the Vorþing, to be spent on care of the outlaw’s dependents and, if anything was left, on other itnerants. It was legal to kill an outlaw, illegal to feed him, shelter him, or help him to leave Iceland. For somewhat less serious offenses, conviction meant lesser outlawry. A lesser outlaw had the right to leave Iceland[40] and could return in three years. For still less serious offenses the punishment was a fine. If it went unpaid, the penalty was lesser outlawry if the fine was paid at the confiscation court, greater outlawry if it was not.

Prosecution was up to the victim or his survivors. If they and the offender agreed on a settlement, the matter was settled. Most cases in the sagas were settled out of court, usually for money damages, sometimes for lesser outlawry or greater outlawry with permission to leave Iceland. Many were settled by arbitration, including the two most serious conflicts that arose prior to the final period of breakdown in the thirteenth century. Calculations by two different scholars suggest that only about a tenth of cases went to a final judgement by the court.

Where a clash resulted in deaths on both sides they were usually set off against each other, sometimes weighted by how important the man killed was, with the excess paid in wergeld or outlawries. In some cases part of the payment for killing one man was the cancellation of the outlawry imposed on another in a previous conflict. Some passages suggest that the prosecutor who got a man outlawed had the power to later lift the sentence from his outlaw.[41]

Icelandic law distinguished between killing and murder—secret killing. After killing a man, one was obliged to announce the fact immediately:

It is prescribed that where men go only one way from a killing, then the killer is to publish the killing as his work within the next twelve hours; but if he is on mountain or fjord then he must do it within twelve hours of returning. He is to go to the first house where he thinks his life is in no danger on that account and tell one or more men legally resident there and state it in this way: “There was an encounter between us,” he is to state, and (p. 154) name the other man and say where it was. “I publish those wounds as my work and all the injury done to him; I publish wounds if wounds are the outcome and killing if killing is the outcome.”(G1 K87 pp. 153-4)

“It is prescribed that if he publishes it in some other way than now told, then it is deemed murder, with the result that it cannot be claimed that the other man, no matter what offense he may have given, died with forfeit immunity, and no grounds of defense are to be accepted.” (G1 K87 p. 154)

“It is prescribed that if a man murders a man, the penalty is outlawry. And it is murder if a man hides it or conceals the corpse or does not admit it.”(G1 K87 p. 154)

Murder cost the killer the ability to raise legal defenses, such as the fact that his victim was an outlaw or had forfeited his immunity by attacking. Concealed taking in some circumstances had a more severe penalty than open taking.[42] In addition, secret killing (murder) or secret taking (theft) was seen as shameful.[43]



In addition to the legal system based on goði and goðorð, there was a system consisting of groups of households called Hreppur. Unlike the legal system, it was geographically based, consisting of groups of neighbors, at least twenty households in each. Its functions included coordinating summer grazing, providing a system of mutual insurance for the members, allocating responsibility for local orphans and indigents and providing a forum for local disputes. The Hreppur  were self governing, but not much is known about their internal structure.[44]


One possible problem in a system of privately enforced law is the poor or weak being unable to enforce their rights. The Icelandic system dealt with this problem by giving the victim a property right—the right to be reimbursed by the criminal—and making that right transferable. The victim could turn over his case to someone else, either gratis or in return for a consideration.[45] A man who did not have sufficient resources, in wealth or allies, to prosecute a case or enforce a verdict could sell it to another who expected to make a profit in both money and reputation by winning the case and collecting the fine. This meant that an attack on even the poorest victim could lead to eventual punishment. A man might volunteer to take on a case not for money but in order to gain status[46] or because the offender was an enemy he wished to harm.

A second objection is that the powerful could commit crimes with impunity, since nobody would be able to enforce a judgment against them. Where power is sufficiently concentrated this might be true; that was one of the problems which led to the eventual breakdown of the Icelandic legal system in the thirteenth century. But so long as power was reasonably dispersed, as it seems to have been for the first two centuries after the system was established, this was a less serious problem. A man who refused to pay his fines or offer a reasonable settlement  and as a result was outlawed would probably not be supported by as many of his friends as the plaintiff seeking to enforce judgment, since in case of violent conflict his defenders would find themselves legally in the wrong.[47] If the lawbreaker defended himself by force, every injury inflicted on the partisans of the other side would result in another suit and every refusal to settle and pay would pull more people into the coalition against him.

There is a scene in Njal's Saga that provides striking evidence of the stability of the system. Conflict between two groups has become so intense that open fighting threatens to break out during the Althing. A leader of one faction asks a benevolent neutral what he will do for them in case of a fight. He replies that he will draw up his people, armed, on one side. If the leader’s men are losing they can retreat behind his, ending the fight. If they are winning, he will block the other side from the best defensive position available then break up the fight before they kill more men than they can afford.[48]  Even when the system seems so near to breaking down, it is still assumed that every enemy killed must eventually be paid for. Each man killed will have friends and relations who are still neutral and will remain so if and only if the killing is made up for by an appropriate wergeld.

A similar pattern holds even in the final period of breakdown, when coalitions are engaged in a virtual civil war. After it is clear which side has won a battle, most of the losers are given quarter, let go unharmed, with the exception of anyone against whom the winners have a particularly serious grudge. When one side manages a successful surprise attack on the home of the opposing leader they kill him, assuming he is present and does not manage to get away, but they let anyone not killed in the process go, again with the same exception. If he turns out to be absent, they may or may not damage the home but are unlikely to massacre the inhabitants.

There are, I think, two explanations for restraint in what was in other ways a pretty unrestrained conflict. One is that you may want to ally next year with the faction you are currently fighting. Acquiring them as allies may require first paying your debts, including debts for people killed. The other is that the people you have just defeated have relatives, not all of whom are in the enemy coalition. To keep them neutral or, better yet, friendly, it may prove necessary to pay for those you killed.

After one attack in which a home was burned, the attackers eventually agree to pay damages for most of those killed inside it but insist that they owe nothing for one man since they not only offered him the opportunity to come out, they offered to pay him to.[49] It is one of many scenes in the Sturlung sagas in which people eventually agree to pay damages for those they have killed.

In order to be deterred, a potential offender has to know that someone is committed in advance to act against those who wrong his potential victim. In Iceland this was done by a system of existing coalitions, some of them goðorð, some clearly defined groups of friends and relatives. If a member of such a coalition was killed, it was in the interest of the other members to collect wergeld for him even if the cost was more than the amount that would be collected; their own safety depended partly on their reputation for doing so. If the killer was unwilling to pay, it was in their interest to push the case through to outlawry and then do their best to kill the outlaw, even at some risk to themselves.

One difficulty with using fines or damage payments as punishments is that criminals might be judgement proof, unable to pay enough to provide adequate deterrence.  The Icelandic system dealt with this in three ways. First, most offenses were ones for which detection was almost certain,[50] hence the cost imposed on offenders did not have to be increased to compensate for a low probability of paying it. Second, the society provided effective credit arrangements. The same coalitions mentioned above provided their members with money to pay large fines.[51] Third, a person unable to discharge his financial obligation could be reduced to a state of temporary slavery until he had worked off his debt. And criminals had a strong incentive to find some way of paying for their crimes, since the alternative was outlawry.

IV: Causes of the Breakdown of the System

One cause seems to have been increased concentration of wealth. and hence power. By the Sturlung period, there were many areas where all or most of the godord were held by one family, reducing or eliminating the ability of the individual thingman to choose his godi and creating a de facto, if imperfect, form of territorial sovereignty.

Why did power become more concentrated? Possibly because a coalition controlling multiple godord could raise a larger force in support of legal claims. In addition, since a coalition controlling multiple godord in the same area would be subject to less competition for thingmen, it would be able to get them on more favorable terms. Those advantages would be balanced, as in the Somali case, by the increased problems of internal conflict within a larger coalition. The Sturlung period is so called after the Sturlung family, which played a large role in its conflicts, frequently with one leading member fighting against another.

Another possible source of concentration of wealth and power was the introduction of Christianity. Churches were under the control of the local landowners who had donated their land and paid for their construction. Individual farmers owed tithes to the church, some fraction of which was likely to end up in the hands of their owners.[52] That provided the equivalent of taxation, a source of revenue to support the power of ambitious men.[53]

A second and related cause of the breakdown was the introduction into Iceland of a foreign ideology--monarchy. By the end of the Sturlung period the chieftains were no longer fighting over who owed what damages to whom but over who should eventually rule Iceland. The Norwegian king gave at least one of the Icelandic leaders the title of Jarl,[54] along with instructions to take over Iceland on his behalf and collect tribute. By the end of the period many of the godord had been transferred by their owners to the king.  Several of the leading figures, when out of Iceland, usually as a result of a settlement that included temporary outlawry, became retainers of the king, in principle obligated to obey his orders. One of them, Snorri Sturluson, was killed in Iceland on the king’s orders in punishment for returning to Iceland without royal permission.


How well the Icelandic institutions worked is a matter of controversy; the sagas are perceived by many as portraying a violent and unjust society tormented by constant feuding. It is difficult to tell whether such judgments are correct. They were written down during or after the Sturlung period, the final violent breakdown of the Icelandic system in the thirteenth century. Their authors may have projected elements of what they saw around them on the earlier periods they described. Also, violence has always been good entertainment, and the saga writers may have selected their material accordingly. Even in a relatively peaceful society novelists might be able to find, over the course of three hundred years, enough conflict for a considerable body of literature. At one point in Njalsaga two successive chapters both have violence in them, which sounds like a lot until you notice that one of the people who participates in the second as an adult had not yet been born in the first.

The quality of violence, in contrast to other medieval literature, is small in scale, intensely personal (every casualty is usually named), and relatively straightforward. Rape and torture are uncommon,[55] the killing of women almost unheard of;[56] in the rare cases when an attacker burns the defender's home, women, children, and servants are first offered an opportunity to leave. According to a calculation by a scholar who went through the Sturlung sagas counting bodies, during more than fifty years of the violent breakdown of the traditional system the number of people killed or executed each year, on a per capita basis, was be roughly equal to the rate of murder and nonnegligent manslaughter in the United States in 1976.[57]

APPENDIX A Wages and Wergelds

Two different monies were in common use in medieval Iceland—silver and vaðmal, woolen cloth. Silver was measured in ounces (aurar) and in marks; the mark contained eight ounces. Wadmal was of a standard width of about a meter and measured in Icelandic ells (alnar) of about 56 centimeters.[58] The value of an ounce (eyrir) of silver varied, during the twelfth and thirteenth centuries, between 6 and 7 1/2 ells.[59] The "law ounce" was set at 6 ells;[60] this appears to have been a money of account, not an attempt at price fixing.

Gragas contains a passage setting maximum wages,[61] possibly an attempt to enforce a monopsonistic cartel agreement by the landowning thingmen against their employees. Porkell Johannesson estimates from the passage that the farm laborer's wage, net of room and board, amounted to about one mark of silver a year and cites another writer who estimated it at about three-quarters of a mark. Johannesson also states that wages net of room and board seem to have been low or zero at the time of settlement but to have risen somewhat by the second half of the tenth century.

These figures give us only a very approximate idea of Icelandic wages. The existence of maximum wage legislation suggests that the equilibrium wage was higher than the legislated wage. But wages, as Þorkell Johannesson points out, must have varied considerably with good and bad years; the legislation might be an attempt to hold wages in good years to a level below the good year equilibrium but above the average wage.

For a second estimate of wages I have taken advantage of the fact that one of the two monetary commodities was woolen cloth, a material which is highly labor intensive. If we knew how many hours went into spinning and weaving an ell of wadmal, we could estimate an upper bound on the market wage rate; if it takes y hours to produce one ell, then the wage of the women making cloth, including the value of any payment in kind they receive, should be less than l/y.

I estimated y in two ways—from figures given by Hoffman for the productivity of Icelandic weavers using the same technology at later periods[62] and from estimates given me by Geraldine Duncan, who has herself worked with a warp- weighted loom and a drop spindle, the tools used by medieval Icelandic weavers.[63] Both methods lead to imprecise results: the first because reports disagree and also because the sources are vague whether the time given is for weaving only or for both weaving and spinning, the second because Mrs. Duncan did not know the precise characteristics of wadmal or how the skill of medieval Icelandic spinners and weavers compared with her own. My conclusion is that it took about a day to spin and weave an ell of wadmal; this estimate could easily be off by a factor of two in either direction. If we assume that, in a relatively poor society such as Iceland, a considerable portion of the income of an ordinary worker went for room and board, this figure is consistent with that given in Gragas.

A rough check on these estimates of wages is provided by the fact that the logsogumadr received an annual salary of 240 ells of wadmal[64] plus a part of the fines for certain minor offenses. While his position was not a full-time one, it involved more than just the two weeks of the Althing; he was required to give information on the law to all comers. Since the man chosen for the post was an unusually talented individual, it does not seem unreasonable that the fixed part of his salary (which, unlike the wages discussed before, did not include room and board) amounted to five year's wages or an amount of wadmal which would have taken about ten months to produce. Thus this figure is not inconsistent with my previous estimate of wages.

During the Sturlung period, when wealth had become relatively concentrated, the richest men had a net worth of about three to four hundred year's production of wadmal or about a thousand cows. The former figure would correspond today to about six million dollars, but the latter to only a few hundred thousand—wages having risen considerably more, over the last millenium, than the price of cattle.

Table 1 gives values for a number of things in ounces, ells, years of production of wadmal, and years of wages. The ounce is assumed to be worth six ells, the year’s production of wadmal to be three hundred ells (three hundred days at one ell/day) and the year's wage to be one mark of forty-eight ells.




Years Production of Wadmal

Years Wages


Normal Price of Male thrall





Carl O. Williams, supra  note 4, at 29

Manumission price of thrall





Sveinbjorn Johnson, supra  note 4, at 225

Wergeld for thrall






Wergeld for free man





Njal's saga, supra  note 3, at 108

Wergeld for free man a






Wergeld for important man





Id.  at 255 ns.

Wergeld for important mana






Law-speaker salary




Vigfusson & Powell, supra  note 1, at 348

Wealth of very rich man
(Sturlung Period)




Einar Olafur Sveinsson, supra  note 44, at 45

Wealth of very rich man
(Sturlung period)





Price of cow
(c. A.D. 1200)




Id.  at 56

a Magnusson and Palsson (Njal's saga, supra note 3, at 63, trans. n.) interpret the ounce by which compensations are measured as probably meaning "an ounce of unrefined silver ... worth four legal ounces," Williams, supra note 4, at 31, interprets it as the legal ounce.

Wergeld for a thrall, the price of a thrall. and the manumission price of a thrall were all equal, as might be expected. The price of a thrall presumably represents the capitalized value of his production net of room and board. It seems at first surprising that this should amount to only a year and a half of wages (also net of room and board), but we must remember that wages, according to Thorkell Johanneson, were lower in the early period, when thralldom was common; thralldom disappeared in Iceland by the early twelfth century and the figures in Gragas may date from later.

The wergeld for a thrall was much lower than for a free man. The value of a thrall to his master would be the capitalized value of his net product. But the value of a free man to himself and his family includes not only his net product but also the value to him of being alive. Food and board are expenses to the owner of a thrall but consumption to a free man. And the costs of the thrall to the owner would include costs of guarding and supervision that would not apply to the free man's calculation of his own value. In one passage in Njalsaga, a trusted household member, presumably a thrall, asks Njal to promise that if he is killed he will be paid for at a free man’s price and Njal agrees.[65]

If we interpret the ounce of Njal's Saga as a legal ounce, the usual wergelds for free men again seem somewhat low, ranging from 12 l/2 year's wages for an ordinary man to twice that for a man of some importance.[66] Here again. we must remember that there is considerable uncertainty in our wage figures. Twelve and a half years' wages might be a reasonable estimate of the value of a man to his family, assuming a market interest rate of between 5 and 10 percent, but it hardly seems to include much allowance for his value to himself. If we accept the interpretation in Magnusson and Palsson[67] of the ounce in which the wergelds of Njal's Saga are paid as an ounce of unrefined silver, worth four legal ounces, the figures seem more reasonable.





StSI is volume I of the translation of Sturlungasaga. StS II is volume II. GI is volume I of the translation of Gragas, GII is volume I. Quote from Gragas are labeled by both section, such as K ŗ 24, and page in the translation.

[1] Gary Becker & George Stigler, Law Enforcement. Malfeasance, and Compensation of Enforcers, 3 J. Legal Stud. I (1975).

[2] William M. Landes & Richard A. Posner, The Private Enforcement of Law, 4 J. Legal Stud. I (1975).

[3] That led to the interesting question of why some offenses are treated as torts, to be prosecuted by the victim, some as crimes, to be prosecuted by the state.. Readers interested in my views of that question will find them in Chapter 18 of my Law’s Order.

[4] "Efficient Institutions for the Private Enforcement of Law." Journal of Legal Studies, June (1984).

[5] Friedman, David "Private Creation and Enforcement of Law -- A Historical Case," Journal of Legal Studies, (March 1979), pp. 399-415.

[6] Some sagas are retellings of material from elsewhere, such as Volsungasaga, the Icelandic version of the German Niebelungenlied. Others are accounts of events in Norway such as Heimskringla, Snorri Sturluson’s saga history of the Norwegian kings. Since they provide little information on Icelandic law I did not include them. That left about 2500 pages of The Complete Sagas of Icelanders, set mostly in Iceland, covering events from the settlement in 870 into the eleventh century, plus nearly a thousand pages of Sturlungasaga, the collection of sagas focussed on the final period of breakdown in the thirteenth century and the events that led up to it.   

[7] I mistakenly believed that a killing only led to outlawry if the killer was unwilling to pay wergeld. That describes what usually happens in the sagas as a result of an out of court settlement but not the legal rule. A successful prosecution resulted in outlawry even if the defendant was willing to pay. I further believed that an outlaw had a period in which he was free to leave Iceland before it was legal to kill him. That again was not a correct description of the legal rule. Once the confiscation court had been held on an outlaw’s property he was fair game and it was illegal to assist him to leave the Island. But it does describe the terms of full outlawry often agreed to in settlements, as well as the terms of lesser outlawry.

[8] Íslandinga saga was written by Sturla Þórðarson, an active participant in many of the events it describes. Prestssaga Guðmundar góða was written by Lambkár Þorgilsson, friend and secretary to Bishop Guðmund, its central figure.

[9] “For all the killings I have now told and also for major wounds men are not to settle without prior leave of the General Assembly. Lesser outlawry is the penalty if men settle in cases which ought not to be settled without prior leave.” Gragas, Volume I, p. 174, K98

One passage in Gragas can be read as implying that a settlement was approved if nobody with a seat on the Lögrétta, which is to say no goði, disapproved of it. But the passage goes on to describe procedures for providing temporary replacements for goðar absent from the Lögrétta, which implies that it is referring to events at the Althing. If any goði had the right to veto any settlement, one would expect some such incidents to appear in the sagas. [K 117 pp. 212-13]

Sigfusson classified cases in a group of sagas according to how they were resolved and found that only about ten percent were settled in court (Sigfusson 160-161). He cites a calculation by Andreas Heusler covering all cases in the sagas of the Icelanders with a similar result. (Sigfusson 154)

[10] [K86, p. 145, 148, 149] [K99, pp. 174-5] [K 105 p. 181 if at assembly must leave]. According to another rule, someone charged with an offense who attends an assembly can escape punishment by showing that he was charged not because he was guilty but in order to keep out of the assembly. [K99 p. 175] That parallels an Athenian case mentioned in Chapter XXX.

[11] StSII 161-2 The Saga of Guðmund Dýri Ingimund, who has killed Helgi, is then himself killed. The settlement of the case sets off Ingimund’s death and the fine on Eyjólf for harboring a murderer against the murder of Helgi. Since Ingimund has not been tried, this appears to support the rule in Gragas. On the other hand, the fact that damages are owed for killing Ingimund appears inconsistent with the Gragas account of forfeit immunity. In V: 14 People of Laxardal Thorolf has killed Hall and asks his kinswoman Vigdis for protection. She replies that “anyone who offers you protection does so at the risk of his own life and property, because such powerful men will be on your trail.” The implication is that sheltering a killer who has not yet been tried can get one into trouble, but it is unclear whether that is a matter of law or only power.

[12] FS: I 279 The Saga of Bjorn, Champion of the Hitardal People. III 85 Njal's Saga, III 246 The Saga of Finnbogi the Mighty

[13] Njal Saga III 78-9 (check)

[14] IV 281-2 The saga of  the People of Reykjadal and of Killer-Skuta, V 130: Bolli Bollason’s tale

[15] Quote the parallel bit from Bedouin practice, from the story of how the long feud was ended.

II 429 Viglund’s saga. III 68 Njal’s saga, III 86 Njal’s saga: 86, III 286 The Saga of the People of Floi,

[16] According to Gragas, GII p. 70 K156, “If a man lies with a slave-woman, he is under penalty for three marks for that, …” (or p. 48?)

[17] The father was a prominent landowner. The couple lived together openly, unable to marry because their relationship was within the very broad reach of the incest prohibition.

[18] The Saga of the Icelanders Pp. 189-194 Þorvald

[19] “The principal in a case [of intercourse] is first the woman's husband if he is born a lawful heir. Then her father. Then a son born a lawful heir, sixteen winters old or older. …” K156 p. 48/70. Possibly the father could have prosecuted but chose not to, his daughter having become a mistress with his consent.

[20] Pregnancy leading to legal charges is mentioned in Njalsaga III: 76, People of Laxardal V 81, StSI: 64 Hvamm-Sturla, The Saga of the Icelanders StSI:238. Pregnancy apparently not leading to legal charges in II 443 Cairn Dweller and IV 395 People of Fljotsdal. Affairs not leading to legal charges in II 354 Sworn Brothers, 120 The Saga of the Icelanders. There is one case where a man volunteers to pay damages to the husband of the woman he has been sleeping with (StSII 168-9 The Saga of Guðmund Dýri) and one case where a man is said to owe compensation for kisses (FS: I:219 Kormak). There is also one case where a man is flirting with a woman and writes her love verses. Her father takes him to court, but the proceedings are violently interrupted. IV 50 People of Vatnsdal

[21]The so-called book-prose theory of saga studies looked upon the Icelandic family sagas basically as fiction, which made historians of early Iceland heavily dependent on the law code, Grágás, for everything that could not be based on archaeological evidence. Over the last two decades, scholars however, have begun using of the family sagas historically again, by supporting their conclusions drawn from them with anthropological studies about remote parts of the earth.” Gunnar Karlsson, Review of Chieftains and Power, Scandinavian Studies pp. 88-89.

[22] Snorri Sturluson, author of Heimskringla and a major figure in the Sturlung period conflicts, comments in the introduction that the most reliable source of historical

[23] Jesse Byock, “Egils Bones,” Scientific American, January 1995, Volume 272 #1 pp. 82-87.

[24] Byock, …

[25] Sigurdsson, Chieftains and Power.

[26] K 117 p. 213

[27] “It is doubtful that Gragas represents an official collection of laws. It should rather be considered either a private collection of laws adopted by the Law Council and written down by individuals, or a ‘rights’ book’ containing notes on legal provisions that may not necessarily have been adopted by law. There is no indication that Gragas was ever consulted as an authority at the general assembly.”

(Review of Jón Viðar Sigurðsson, Chieftains and Power, by Gunnar Karlsson. Scandinavian Studies volume 73 no. 1 pp. 88-89)

[28] V 350 Hrumund the Lame

[29] G1 K 28 p. 53. It’s possible that bringing an army to the Althing does not count as bringing men to court, since the court is only part of what is happening at the Althing.

[30] Some estimates put it at about 10%.

[31] Barthi Guthmundsson, The Origin of the Icelanders (Lee M. Hollander trans. 1967), argues that the settlers were in large part Danes who had colonized in Norway and thus brought Danish institutions with them to Iceland.

[32] There are passages in the sagas which suggest that the godord may have been viewed as a territorial authority in the very early years of the system. The Saga of Gunnar, The Fool of Keldugnap, III 435 (check)

[33] The twelve goðar in the North Quarter appointed only nine judges to the quarter courts. My description of the system is that accepted by most scholars. Jón Viðar Sigurðsson argues that it is a back projection from later accounts, that the actual system in the early period was considerably less tidy.

[34] The Icelandic judges correspond more nearly to the jurymen of our system, since it was up to them to determine guilt or innocence. There was no equivalent of our judge, although experts in the law could be consulted by the court.

[35] C. A Vansittart Conybeare, The Place of Iceland in the History of European Institutions 48 (1877).

[36] While this is the generally accepted view, some have argued that the judges for a quarter court were appointed by the goðar from that quarter. Sveinbjorn Johnson, supra  note 4, at 64; and James Bryce, Studies in History and Jurisprudence 274 (1901),

[37] While there was nothing strictly equivalent to our system of appeals, claims that a case had been handled illegally in one court could be resolved in a higher court. In a famous case in Njalssaga the defendant tricks the prosecution into prosecuting him in the wrong court by secretly changing his godord, and hence his quarter, in order to be able to sue the prosecutors in the fifth court for doing so. Id. at 93-94; Njal's Saga, supra  note 3, at 309-310.

[38] “It is only lawful for a man to have an assembly attachment in a different Quarter from the one he lives in if the chieftain in question is permitted at Lögberg to accept someone from outside the Quarter as a man of his assembly third.G1 K83 p. 141.” This rule appears inconsistent with the trick in Njalsaga (Sagas of Icelanders V. III p. 179), where Glosi transfers his goðorð to his brother and becomes the thingman of a goði in a different quarter, changing what court he should be sued in.  A man could declare what goði he was the thingman of at the Althing or the Vorþing.

[39]  “Men shall pay assembly attendance dues at the rate they agree on with the chieftain in each assembly third.” (G1 K23 p. 44).

[40] He was immune from killing until he left, provided that he remained within a limited area assigned to him.

[41] StSII: 100 “Þorkel Flosason, declared outlaw the previous summer, paid a visit to Þorvarð and, placing his head on the table before him, awaited his decision. Þorvarð granted him his parden and told him to go in peace wherever he wished”

[42] Full outlawry (I, 246) was the penalty for stealing food, irrespective of its value. It was also the penalty for stealing anything else worth more than half an ounce-unit, given that any attempt was made to conceal the theft. Jewish law also penalized theft more severely than robbery. Maimonides

[43] References to Njalsaga (Halgerd’s theft) and Egilsaga.

[44] Jesse Byock, Viking Age Iceland, pp. 137-8.

[45] For examples, see Njal Saga V III p. 27, 76,

[46] As happens in Hrafnkel Saga. The same pattern appears in the Commanche system, as mentioned in Chapter XXX p.

[47] Rominchal reference

[48] Njal Saga V III about p. 195 check.

[49] Sturlungasaga Volume II p. 184

[50] Since concealing the crime was considered shameful and if unsuccessful imposed additional legal consequences.

[51] In the Somali system described in Chapter XXX the division of payments was defined in advance by the contract among members of a coalition. The Icelandic system was informal, but there are cases in the sagas where a large fine imposed on a leader in a settlement is paid largely by voluntary donations by his followers. Gragas refers in several places to life-ring payments which seem to be from various kin of the offender to the corresponding kin of the victim, but they do not seem to appear in the sagas.

[52] Each thing-tax paying farmer owed one percent of his wealth each year as tithe. A quarter went to the Hreppur for maintainance of the poor. A quarter went to the bishop. A quarter went to the person who controlled the church for its maintainance, a quarter to the priest—who might be the owner, a member of his household, or one of his servants. Thus as much as half of the tithe could end up going to the landowner who had provided the land for the church.

[53] Byock 328-9.

[54] Gizur jarl. According to some accounts, Snorri Stuluson was also given the title.

[55] There are examples of torture in Hrafnkel saga . In one of the Sturlung sagas (check which one) Sturla tricks Órĺkja in his power and said to have been blinded and gelded. But later in the saga he appears to still be able to see, a fact never explained. Possibly the person actually responsible for his mutilation only pretended to do it. One incident involves rape or intended rape of a man’s wife and daughter, another carrying off of a woman with the intention of marrying her and then returning her when it becomes clear she cannot be pursuaded. (Sturlungsaga vol. 1 pp. 207-8)

[56] The only cases I have come across where a woman is deliberately killed involve the execution of witches. There are several incidents in the sagas where a woman attacks a man, in one case trying to stab him with the sword of her brother, who he has helped to kill, in another striking a man across the face with the purse of silver with which he has tried to bribe her to betray her outlawed (brother? Check). In both of those, no action is taken in retaliation.

[57] Id at 72 gives an estimate of three hundred and fifty killed in battle or executed during a fifty-two-year period (1208-12601). The population of Iceland was about seventy thousand. For the U.S. figures, see Michael S. Hindelang et. al., Sourcebook of Criminal Justice Statistics-- 1976, at 443 (1977). Einar Olafur Sveinsson. The Age of the Sturlungs 68, 73 (Johann S. Hannesson trans. 953) (Islandica vol. 36).

[58] Marta Hoffman, The Warp-Weighted Loom 213 (1964).

[59] Knut Gjerset, History of Iceland 206 (1924).

[60] Njal's Saga 41, trans. n. Also Þorkell Johannesson, Die Stellung der Freien Arbeiter in Island 37 (1933).

[61] Find gragas cite

[62] Marta Hoffman, The Warp-Weighted Loom  (1964) 215-16.

[63] Private communication

[64] G1 K 116 p. 209.

[65] Njal saga cite.

[66] In comparing this figure with current sentencing levels for murder or manslaughter, one must remember that killing, in Icelandic law, was distinguished from murder by the fact that the killer "turned himself in." Thus even if the average sentence served by the convicted killer~ in our society were as high as 12 1/2 years--which it surely is not--the corresponding expected punishment would be much higher in the Icelandic case.

[67] Njal's Saga, supra  note 3, at 63.[check if the reference is in the translation I’m now using.]