3
The Society

Introduction

Going from the analysis of conflicts to society as a whole, we take a step away from Snorri's explicit ideas to his basic assumptions based on a common mentalité . Although the conflicts are his main interest, Snorri is no sociologist and does not attempt to analyze Norwegian society. In his narrative, however, he inevitably reveals some kind of social organization, though we must have in mind that his exclusive interest in conflicts and dramatic events may distort the picture of the normal functioning of society.

In studying society on the basis of Snorri's narrative, we have to consider two kinds of social organization: first, the groups that are opposed to one another in the struggle for power, and second, the way in which individuals are ranked in the social hierarchy. In modern industrial societies we expect some kind of correspondence between the two but this need not be the case in more traditional societies. Let us start with the first kind of organization.

The Factions in the Conflicts: Family Clans or Ad Hoc Groups?

The discussion in the preceding pages has led to the conclusion that Snorri's conflicts are mostly feuds, not conflicts over ideology, constitutional principles, or class interest, and further that the feuds, though

often sparked by accidental clashes, usually involve conflicting long-term interests between the actors.

Though the conflicts may be termed personal rather than ideological or social, it remains to consider more closely the formation of factions and consequently the lines of division within Snorri's society. Norwegian and Icelandic society of the eariler Middle Ages is often referred to as a "society of kindreds," without a very precise definition of the term (e.g., Johnsen, 1948; Andersen, 1977: 70 f., 185 f., 247 f., 336 ff.). By contrast, authors who have attempted a more exact analysis of the kinship system have generally concluded that it was bilateral and thus not likely to form the basis of extended clans and that it was fairly early surpassed by other forms of social organization. This impression is based on comparative analyses of the laws, but no examination of the whole material according to modern anthropological principles has been carried out. Above all, the sagas and documentary sources have been relatively neglected.^[1] It is evidently outside the scope of the present work to analyse the kinship system in full. I shall confine myself to Snorri and to the political relevance of the system.

The Importance of Kinship

The kindred (ætt) is frequently mentioned in Heimskringla . Relatives can give important support. When a man is killed, it is the duty of his relatives to take revenge or seek compensation. It is a laudable act to perform this duty and dishonorable to shrink from it.^[2] Conversely, one should not kill one's own relatives. The English Earl Tósti even declares that he prefers to be killed by his brother Harold Godwineson rather than to kill him, despite the fact that they are enemies (HHarð . chap. 91). Haraldr gilli's treatment of his nephew Magnús blindi, which seems to us an example of the utmost cruelty, may be explained in this way. Magnús was blinded, castrated, and had one of his feet cut off. The purpose of this is clearly stated: he should not be able to be king any more (MB.HG chap. 8). Being unable to see or to walk in a normal way, he cannot lead an army. As a castrate, he cannot have offspring, apart from the fact that he has lost his power as a man and is thus hardly fit to be a leader of other men. Magnús's fate may appear worse than death and be considered—as has often been done—an act of sadism. Though there are indications that people thought like this in the Middle Ages as well,^[3] the aftermath of the story shows that he was still dangerous and thus that the most prudent thing would have been to kill him—if there were no moral objections to it.^[4] Snorri's story of King Hroerekr, who was also blinded but whom St. Óláfr did not want to kill, seems to confirm that this was the case.^[5] An Icelandic example points in the same

direction: Sturla Sigvatsson in 1236 tried and partly succeeded in blinding and castrating his cousin Óroekja, Snorri's son, but did not attempt to kill him (Sturl . I: 485; see also Heusler, 1912: 36). Though there are examples of near relatives being killed, this mostly happens in battle and not at the direct order of their kinsmen. This applies among others to Ingi's brothers Sigurðr and Eysteinn. The former was killed by arrows in the battle in Bergen following the conflict between him and Ingi (Ingi chap. 28), whereas the latter was taken captive by a former adherent and killed (Ingi chap. 32). Though this was hardly against Ingi's will, he could not be blamed directly for carrying out or ordering the execution. King Eiríkr blóðøx, who killed two of his brothers (HHárf . chaps. 35-36) is a more direct exception to the rule. But Eiríkr was an extremely cruel and violent man, whose acts Snorri describes with some disapproval. By contrast, Erlingr skakki, who was a hard man and ruthlessly exterminated all potential rivals of his son Magnús, is no real exception, because none of the men he killed were his blood relations.^[6]

How far do these ties of kinship extend? According to the laws, both the Norwegian and Icelandic ones were fairly extensive, but the practical importance of the rules there may be doubted. According to Gaunt (1983: 206), who quotes Johnsen (1948: 76 ff.), there are few examples in the sagas of revenge for other than very close relatives, such as fathers or brothers. This is not quite the impression we receive from Heimskringla , which does give examples of revenge or attempts at revenge for more distant relatives. When Haraldr harðráði has killed Einarr þambarskelfir and his son Eindriði, he fears the revenge of Hákon Ívarsson, the grandnephew of Einarr's wife, whose father was thus Eindriði's cousin. Þórir hundr revenges his nephew Asbjorn^[*] selsbani (OH chap. 133). Erlingr Skjálgsson, who risks a major conflict with the king for Asbjorn's sake, is also his uncle— Asbjorn's mother is Erlingr's sister (OH chap. 117: 242). In some other cases, the degree of kinship is not indicated. Erlingr Skjálgsson is revenged by a man called Vígleikr Árnason, of whom nothing else is known (OH chap. 178). It has been suggested that the duty of revenge did not depend on the degree of kinship but that it pertained to the heir (Sommerfelt, 1974: 144 f). This fits in well with the example of Hákon Ívarsson.^[7]

Whether or not to take revenge, however, is hardly a question of duties only. Individual interests also come into consideration. For a king or magnate to abstain from revenge without compensation is clearly out of the question, as this normally means loss of honor and consequently of political influence. There are indications both in the laws and the Icelandic sagas that it is considered more honorable to take revenge than to receive compensation.^[8] However, these alternatives are not strictly opposed; there is a sort of sliding transition between them: if the com-

pensation is sufficient, it may be an honorable alternative. This is expressed in the institution of sjalfdoemi , that is, that the offended party becomes judge in his own case. Economically, he may seek any compensation he wants, and politically, he has forced his adversary to total capitulation.^[9]

The negotiations between King Haraldr harðráði and Hákon Ívarsson after Einarr þambarskelfir's death at Haraldr's hands may serve as an example (HHarð . chaps. 45-47). When Finnr Árnason as the king's representative offers compensation in case Hákon is willing to abstain from revenge, Hákon first points to his duty to revenge his kinsman Eindriði, Einarr's son. Finnr tempts him with a large compensation—he may ask the terms he likes—and threatens him with punishment or dishonor for treason against the king. Hákon then asks for the daughter of King Magnús góði in marriage, thus seizing the opportunity that the king's offense offers him to further his own interests.

Hákon's willingness to reach an agreement may be explained either from the fact that it was a question of a relatively distant relative or from political reasons, or—of course—both. The way Snorri tells the story points to the political reasons as the more important, but we cannot exclude the possibility that the relevance of the degree of kinship was evident to Snorri's readers and thus implied. As Snorri is mainly interested in conflicts and therefore rarely quotes examples of peaceful settlements, general rules of behavior in such cases, including the importance of the degree of kinship for the course of action that was taken, can hardly be reconstructed from his examples. Generally, given the importance of individual interests in Snorri, he probably regards them as more important than the degree of kinship. Further, the story of Magnús góði shows the conflict between political considerations and the wish for revenge. When Magnús grows up, he persecutes his father's enemies, helping Ásmundr Grankellsson to kill Hárekr at Tjøtta (MG chap. 12), forcing Kálfr Árnason to leave the country (MG chap. 14), and confiscating the property of many men who had fought against Óláfr in the battle of Stiklestad (MG chap. 15). This leads to opposition against him, and he mends his ways and becomes the good king Magnús (chap. 16), though Kálfr remains in exile till after his death (HHarð . chap. 51) and Snorri does not mention what happens to the confiscated property. Though Snorri is fairly vague on this point, the story clearly suggests that it is necessary for a king to show moderation in carrying out revenge.

Although there seems to be a certain logic in the duty of revenge being coextensive with the right to inherit, the ban against killing relatives may well have had a wider application. To abstain from killing is clearly less to ask than the active solidarity implied in the duty to take revenge. Erlingr Skjálgsson was quite a distant relative of Áslákr—they

were second cousins (OT chap. 54)—but Áslákr is nevertheless blamed for killing him. In the case of St. Óláfr and King Hroerekr, Snorri confines himself to stating that they were both descended from Haraldr hárfagri, which probably means that Óláfr's father and Hroerekr were second cousins.^[10] The kinship between Óláfr Tryggvason and Hárekr at Tjøtta, which is one of the reasons for Óláfr to release Hárekr (OT chap. 75: 392), was equally distant.^[11] There are thus at least some indications of quite extensive kinship ties in Snorri's society. Their practical and political importance, however, is another matter. Halvdan Koht tried to demonstrate the importance of kinship for faction formation during the inner conflicts from the tenth century until the first half of the thirteenth century (1921: 114 ff.; 1936: 89-104). He managed to show that the divisions that Edvard Bull (1917: 177 ff.) had regarded as the result of regional differences were better explained by kinship or marriage alliances (see Bagge, 1986: 160 ff.). His attempt to explain the political conflicts through a long-term opposition between kindreds was less successful, however, because he did not cross-check his material by examining (1) how many equally near relatives were on the opposite side; and (2) how many adherents of the respective factions had family backgrounds that were irrevelant to their choice of sides. Nor did he distinguish between kinsmen and in-laws. Admittedly, marriage probably created strong ties. But since marriage is the result of a choice, faction adherence may equally well be its cause as its effect. Finally, in analyzing Snorri, we may note that he does not always state the genealogies explicitly, so that they have to be reconstructed laboriously, a fact that indicates that Snorri was not particularly concerned with kinship ties as the explanation to faction loyalties.

More specifically, the following arguments may be adduced against Koht's results. Koht attaches great importance to the generation-long conflict between Haraldr hárfagri's kin and that of the earls of Lade, which is quite correct. He can also show that many of the magnates who rebelled against St. Óláfr and his successors were related by blood or marriage to the earls. But some of them were related to the kings as well, such as Hroerekr and Hárekr at Tjøtta. Moreover, many of them were not consistent enemies of the kings, as they ought to have been if the conflicts were along kinship lines. Men like Einarr þambarskelfir, Kálfr Árnason, and to some extent Finnr Árnason were power politicians who changed sides according to their interests, not consistent opponents of the royal family.^[12] Further, the inner conflict until the time of Haraldr harðráði cannot be summarized as a conflict between the royal line and that of the earls of Lade. First, the two kins were themselves related. King Haraldr hárfagri married one of his daughters to Earl Hákon Grjótgarðsson of Lade, whereas the earl's son, Sigurðr, married

one of Haraldr's granddaughters. Earl Hákon Sigurðarson was thus Haraldr's great-grandson, descending from him both through his mother and father (HHárf . chaps. 9, 37; HG chap. 11), and equally related to the royal line as Hárekr at Tjøtta to the earls of Lade. Second, the conflict was as much between members of the royal family as between the kings and the earls. Nor was there perfect unity and harmony within the earls' line: Grjótgarðr betrayed his elder brother and rival Sigurðr to the Eiríkssons and was in turn killed in battle by his nephew Hákon (HGráf . chaps. 4-5, 15). There are also other examples of relatives belonging to different factions or rivalry between them.^[13] As for the connection between this period and the civil wars of the twelfth century, the fact that both Grégóríús and Erlingr skakki were probably descended from the earls of Lade is less significant than the former's close friendship with King Ingi and the latter's marriage to the daughter of King Sigurðr jórsalafari (see Koht, 1921: 118 ff.). Though kinship ties were not unimportant, there is thus slight evidence for faction divisions being consistently determined by them. It was probably more usual that relatives belonged to the same faction than to different ones, but exceptions to this were sufficiently numerous to make it difficult to regard the factions simply as family clans.

Rather, we have to make a fundamental distinction between particular duties between kinsmen on the one hand and active solidarity and faction formation on the other, that is between grid and group (see Douglas, 1970: viii ff.). In a bilateral kinship system such as the Norwegian and Icelandic ones, only very close relatives, that is, siblings of the same parents, belong to exactly the same kindred. Consequently, there is no rule that determines the loyalty of an individual in a conflict between two of his kinsmen, who may themselves be unrelated to one another. Thus, though there may be strong solidarity within the smaller family unit, larger factions are likely to be formed through ties other than those of kinship, or individual interests may determine the choice (see n. 1).

The actors in the political game very often seem to be heads of households who are supported by sons and brothers. Erlingr Skjálgsson's sons play an important part in his struggles with St. Óláfr, among other things in the story of Asbjorn^[*] selsbani. Þorbergr Árnason turns to his brothers for help in his conflict with the king. The role of women can also be regarded in this context. To Snorri, women are not normally actors in the political game, as they do not take part in war. But they may play an important indirect role, by influencing their husbands, sons, brothers, and so forth. Their most important role in Heimskringla as in the saga literature in general is that of inciters. They bring pressure on their reluctant male relatives to take revenge rather than reaching a peaceful

settlement. Examples of this are Þórir hundr's sister-in-law and Kálfr Árnason's wife, both named Sigríðr. There has been some discussion whether this role corresponds to historical reality or whether it is a literary phenomenon, inspired by the negative picture of women in contemporary clerical literature.^[14] Though I do not want to express a definite opinion on this question with only Snorri as a source, this role of women certainly makes sense in the political game as described in Heimskringla . To women, who are not constant participants in the political game, the ties of kinship and the loss of a dear relative are the most important consideration, whereas long-term political interests often make men seek another solution.

However, this is not the only role of women in Heimskringla . They may also act in the interest of peace or to help their male relatives, such as the Swedish Princess Ingigerðr or Kristín Sigurðardóttir, Erlingr skakki's wife. Or they may use their influence to serve their own interests, revenge insults against their honor or improve their position. Sigriðr storráða seeks revenge against Óláfr Tryggvason for his insult against her and gets her husband and son to make war against him and defeat him in the battle of  . At the court of the young King Magnús góði, his mother Álfhildr and the Queen dowager Ástríðr compete for prestige, and the former's behavior serves to illustrate the general "law" that people who come to great power often grow arrogant (MG chap. 7). But women rarely pursue long-term political aims in the same way as the male protagonists in the game. One who does is Queen Gunnhildr, the moving force behind the ruthless policy of the Eiríkssons (see Christophersen, 1987). In one particular context, Queen Ingiríðr, King Ingi's mother, acts in the same way, being one of the "hawks" in the assembly that decides to fight King Sigurðr in Bergen in 1155. And Queen Ástríðr, St. Óláfr's widow, plays an active role to secure the throne for her stepson, Magnús, who is a minor. These are exceptional cases, however. Generally, women are more closely linked to family and kinship in two ways. They normally act through their male relatives and their aim is often, though not always, to strengthen or protect the ties of kinship.^[15]

Patronage and Marriage Ties

Similar ties of loyalty as those between kindreds existed between a king or magnate and his men. The duty of revenging "friends" or lords/ clients is apparently the same as the duty to revenge kinsmen. St. Óláfr's intended execution of Asbjorn^[*] for killing his ármaðr Selþórir is quite explicable in these terms. He also claims compensation for his hirðmaðr Karli, who was killed by Þórir hundr in revenge for Asbjorn, by humiliating Þórir and forcing him to buy his peace for an enormous amount of

gold. King Ingi feels obliged to revenge his friend Grégóríús and gives this as one of his reasons for not withdrawing from battle. The provision in the laws of a special fine for killing the king's men, which in The King's Mirror is pointed out as one of the advantages of entering the king's service (Kgs .: 41, lines 20-21), may be regarded as a formalization of this solidarity between the king and his men. Conversely, it is the duty of hirðmenn or housecarls of great men to revenge their lord, as is evident from the episodes when this duty is not carried out because of the panic caused by the loss of the leader.

Snorri gives no clear indication of the relative importance of the loyalty between a lord and his men and that between kinsmen. As in the case of kinship, he clearly indicates that interests may lead men to break the ties of loyalty to their leader, though the ethical evaluation of this depends on social status. The ties of loyalty between a king or magnate and his men are thus not necessarily stronger than those of kinship. Nevertheless, the former ties are probably more important for faction formation, first, because they do not present the same obstacles to group formation, and second, because they are more likely to serve individual interests. Depending on the distance in power and status between the partners and on benefits bestowed and received, alliances based on friendship, loyalty, or mutual interests could be created that were stronger than those forged by kinship ties.

If we look at the reality behind the apparently chaotic picture Snorri gives of respect and disregard of kinship ties, a certain pattern seems to emerge: kinship ties are weakened by concentration of power. Kinsmen offer valuable protection against aggressors and may be allies in struggles for power and influence, as Snorri repeatedly demonstrates.^[16] But kinsmen are probably more difficult to command than nonrelated clients, at least with the kinship structure that prevailed in medieval Norway and Iceland. Consequently, they become a hindrance rather than an asset to really mighty men. The development of office points in the same direction. Although private property was normally divided between heirs, kingship and the office of lendr maðr were not, at least not in the same sense. Heirs to the kingdom might stick together against nonrelated rivals, like the Eiríkssons, but generally, they were rivals rather than allies.

Within the royal line conflicts not only occur, but seem to be the normal way of things. According to Gurevich, the explanation for this is a hereditary damnation deriving from an episode in the Ynglinga period (Gurevich, 1971: 49 f.; see also Sørensen, 1977: 165). From a modern point of view, the obvious explanation is that kingship is an office, which cannot be divided as easily as private property and, more generally, that the concentration of power in connection with the unification of the

kingdom worked against solidarity between the members of the royal family. It is difficult to tell how far Snorri was aware of these facts. However, though he may have intended the episode in the Ynglinga saga as an explanation when he wrote it, his narrative as a whole and his general analysis of politics suggest that he regarded conflicts over power and resources as a normal phenomenon, even between kinsmen, and did not really need a "mythical" explanation like the one in the Ynglinga saga .

Lower offices, like that of earl or lendr maðr , are in principle treated in the same way as kingship. Only one member of a family is normally appointed, and he is in a superior position compared to his brothers. In this way the Eiríkssons can play off Grjótgarðr against his brother Sigurðr. Þórir hundr is in greater esteem than his brother Sigurðr because he is a lendr maðr , and Sigurðr's son Asbjorn's^[*] hospitality is an attempt to rival him. The case of Erlingr Skjálgsson is particularly instructive. He belongs to the mightiest kindred in Southwest Norway (Rogaland, Hordaland), whose members seem to act together as a clan. Erlingr is chosen to marry King Oláfr Tryggvason's sister. He is appointed lendr maðr by the king, receives great veitslur , and becomes the real ruler of Western Norway, thus changing the collective leadership of his family into a personal principality. Consequently, his kinsman Áslákr fitjaskalli becomes his greatest rival and ultimately kills him. However, rivalry between kinsmen is not the normal way of things within the aristocracy to the same extent as within the royal family. Members of the aristocracy were in greater need of allies on their own social level, and their position in society did not depend on holding an office in the same sense as that of the king.

My last example raises the question of marriage ties, which in feudal Europe seem to have become relatively more important in the high Middle Ages with the change from extended family clans to lineages, and in connection with this the introduction of primogeniture (Duby, 1977: 59 ff., 149 ff;, and 1985: 92 ff., 227 ff.; Freed, 1986: 560 ff.). Such a radical change did not take place in Norway and Iceland, though there were tendencies in this direction.^[17] Nevertheless, marriage seems to have been able to forge important political ties, equally important, if not more, than kinship (see Sigurðsson, 1989: 108 f. on Iceland). This seems likely enough in a bilateral system such as that of Norway and Iceland. Moreover, marriage ties are more easily adapted to new realities, since they can be used to strengthen alliances that are the results of political interests. Marriage ties may therefore well have been strengthened while kinship ties were weakened. In any case, marriage plays a prominent part in Snorri. Alliances between great men are often confirmed through marriage. Haraldr hárfagri's alliance with Earl Hákon Grjót-

garðsson of Lade, which is of fundamental importance for his conquest of Norway, is confirmed by Haraldr marrying the earl's daughter (HHárf . chaps. 7, 9). Earl Hákon Sigurðarson's best friend is his brother-in-law, to whom he also marries his daughter (OT chap. 19). Óláfr Tryggvason uses his two brothers-in-law, together with his maternal uncle and stepfather, to make the people of Viken accept Christianity (OT chap. 53). Erlingr Skjálgsson's marriage to Óláfr Tryggvason's sister brings Western Norway to Óláfr's side while at the same time laying the foundation of Erlingr's greatness. The earls who rule the country after Óláfr's death link the leading men of the country, Erlingr Skjálgsson and Einarr þambarskelfir, to their cause through marriage alliances (OH chaps. 21, 31). The story of St. Óláfr's negotiations with the king of Sweden may also be understood in this perspective. Most of Koht's examples of alliances based on kinship during the civil wars of the twelfth century (1921: 118 f.) are actually marriage alliances, which give the impression that the factions were built up systematically through intermarriage.

The normal pattern of marriage alliances seems to be that if the partners have different rank, the woman has the higher, as is also the case in feudal Europe (see Duby, 1985: 143 f.). The king marries the daughter of another king, whereas the king's daughters or sisters marry magnates within the country or their sons.^[18] Female members of the royal house are thus an important resource in building up a faction around the king, and the alliances formed in this way very often seem to last.^[19] The kings' concubines and illegitimate sons serve to form alliances in a similar way, as does the institution of fostering.^[20] This development is quite logical. On the one hand, politics is based on personal relationships. On the other, the permanent ties of kinship are often opposed to individual interests (see Wolf, 1982: 93 ff.; Verdery, 1988: 268). The logical solution is to create bonds between men that resemble family ties, but which are the result of their deliberate choice. The ties of loyalty between a king or magnate and his men may be regarded in the same way. They imply much the same duties as between kinsmen and have similar emotional overtones, while at the same time being stronger because they take into account individual interests.

The examples quoted above cannot without reservations be considered evidence of an actual historical development. On a priori grounds, however, it seems likely that change along these lines should have taken place, given the apparently rather loose kindred structure at the outset. But as some form of principalities may have existed long before the unification of the kingdom in the late ninth century, the alternatives, solidarity within the kindred or individual assertion, may have presented themselves long before Snorri's time. Nor is there any evidence that

Snorri regarded these alternatives in the perspective of a historical development. However, he obviously did reflect on the problem.

In the discussion between the petty kings of Eastern Norway over St. Óláfr's claim to the throne the kings Hringr and Hroerekr represent diametrically opposed attitudes on one point, namely on kinship. Hringr regards it as a promotion for the whole kindred if one of its members ascends to the throne, whereas Hroe ekr thinks strictly in individual terms, seeing no reasons for supporting Óláfr other than personal gain. Snorri's own point of view does not emerge from the speeches. On the one hand, Óláfr is the great hero of the saga, and Hringr is thus on the right side; on the other hand, Hroerekr is an excellent analyst of politics and, as far as the interests of the petty kings are concerned, ultimately turns out to be right. The most probable conclusion is that Hringr represents the ideal of kindred solidarity, whereas Hroerekr better reflects the way people actually behaved, as is evident in Snorri's numerous examples of rivalry between kindreds and the pursuit of individual glory and power.

Snorri's particular interest in the problem may be explained from personal experience. Snorri lived through the greatest period of power concentration in Iceland and took active part in the process. In these struggles, his own kinsmen were as much his rivals as allies. In the 1230s his own brother and nephew even became his most bitter enemies. Though the Sturlungar may have been exceptional in this respect, it is easy to imagine that the emergence of larger and stronger principalities led by individual magnates might result in increased rivalry within the kindreds.

We may then conclude from the preceding analysis of the conflicts and the present one of the ties of loyalty within the factions that Snorri's society is neither a society based on group or class solidarity, nor a society of kindreds. It is primarily a society of individual kings or magnates competing for power or resources, where clashes or revenge are more pretexts than real reasons for conflicts, and where alliances are the results of personal interests.

Society as a Whole: Conflict or Harmony?

From a modern point of view, the society depicted in Snorri's narrative resembles sheer anarachy. Armed conflicts, mostly between individuals and alliances based on kinship or mutual interests, are quite normal, and they are conducted in a way in which most means are allowed that may lead to victory, the "rules of the game" being few and vaguely defined. Consequently, very little seems to tie this society together. This impression, however, is only partly correct. Snorri focuses

on armed conflicts, whereas the elements that unite the members of society are only mentioned occasionally. First, peaceful means of solving conflicts are clearly underrepresented in the sagas, as they do not lead to important events. Second, in Snorri's society as in many other traditional societies (see Gluckman, 1965: 140 ff.) armed conflicts are not necessarily against the social order. They are part of it and may even serve to preserve it, in a similar way as the political conflicts that are conducted by peaceful means and according to strict rules in modern society.

This sense of the fundamental unity of society despite internal conflicts is expressed in Snorri's comment on the measure Erlingr skakki got passed against his adversary Earl Sigurðr, the leader of the young King Sigurðr Markúsfóstri's faction. After Erlingr's victory over King Hákon herðibreiðr renewed the conflict, but had small support and mainly conducted guerilla warfare, to the detriment of the people of Viken. At a popular assembly Erlingr had his adversary condemned to the devil. Snorri is evidently shocked at such unprecedented harshness and calls it an abominable act (ME chap. 10). The reason for this reaction is probably a fundamental distinction between conflicts "within" society, such as conflicts between pretenders and the—very rare—conflicts when society as a whole must defend its fundamental values. To use the sanctions meant for this last category of enemies against one's normal adversaries is such a horrible break against the rules of the game that it deserves the strongest condemnation in the whole of Heimskringla . By placing his adversaries outside society, Erlingr has broken one of its fundamental rules, the one that unites both parties in a conflict in a fundamental community.^[21] Snorri's reasoning may be compared to the basic idea of modern democratic politics of a consensus concerning the fundamental values of society and politics, despite the constant struggle between different parties.

Apart from such extreme cases, the unity of society is not often focused in Snorri's narrative, probably because it is taken for granted. But Snorri clearly has an idea of Norwegian national unity as opposed to other countries. He frequently expresses nationalistic bias when describing conflicts with other countries and he mentions as a fact that it is usually difficult for a foreigner to be accepted in another country, in particular for a king to conquer another country than his own and be accepted as its ruler. His most important expression of this idea, however, is indirect, consisting in the kind of conflicts that do not occur. As we have seen, a conflict is not a bellum omnium contra omnes , but takes place within particular layers of society, whose members are allowed to claim the positions that may lead to major conflicts. Total victory or defeat has dramatic consequences for the chief participants but does not

change very much. That is: the fundamental order of society is not really affected by the conflicts. The warring factions are united by a common social order and common values, which are implicit. In contrast to what is often the case in modern society and what has often been stated about medieval society, this social order with its inequality and hierarchy is not subject to conflicts but taken for granted, the conflicts taking place between people who belong to approximately the same order, usually at the top.

The Hierarchy of Society

In his analysis of Snorri's view of society, Gudmund Sandvik points out a tripartite structure: king, magnates, people. In terms of political tension or conflict, however, the line of division goes between king and magnates, whereas there is complete solidarity between magnates and people, the two terms often being synonymous. To Sandvik, Snorri's view of society is essentially aristocratic. Though the magnates ultimately have their basis of power in the people, they are not its representatives in the strict sense. They take the political decisions, and the people automatically support them. When their leader is killed, the people are normally helpless even when they are numerically superior. Similarly, discussions in the popular assemblies are solely between magnates or between the magnates and the king, and the people are easily lead by the strongest or most influential person (Sandvik, 1955: 45 ff.).

In its main outlines, this description seems correct. Although Snorri mentions quite a number of social categories, from the king at the top to the slave at the bottom of the social ladder, his main line of division goes between these three. Slaves and freedmen are mentioned, but play no great part, at least not from a political point of view. Though there must have been distinctions of wealth and influence within the people, Snorri treats it as one group, and in fact rarely mentions its individual members. Nor are the magnates clearly distinguished from the people.

The usual Old Norse term for the people, which is also common in Snorri, bóndi , pl. boendr , has no exact English equivalent. It is derived from búa (= to live at or inhabit a place) and is used of ordinary, free men who lived on farms that they owned or rented, whatever their wealth and position in society. It may thus include very prominent men, who were in reality magnates. Þorgnýr  , one of the mightiest men in Sweden and the leader of the opposition against King Óláfr of Sweden, is called bóndi , and even uses the term for himself (OH chap. 80). So is Járnskeggi, the leader of the people in Trøndelag and Einarr þambarskelfir's kinsman (OT chap. 66; see also HGráf . chap. 14). When Ragnhildr, the daughter of the late King Magnús góði, refuses to marry

Hákon Ívarsson, one of the mightiest men in the country, she calls him bóndi (HHarð . chap. 48: 142). The term is thus synonymous with útíginn maðr (a man not holding a tígn , i.e., high rank, usually that of king or earl), which is used both in this and other contexts (ibid.; see also OT chap. 56 on Erlingr Skjálgsson). Nor is the distinction between king and magnates entirely clear. From the time of Haraldr hárfagri there had normally been one or more persons to claim the right to rule the whole country, usually with the title of king. Besides, there were petty kings until the reign of St. Óláfr. As appears from Óláfr's saga, their position was not fundamentally different from that of the mightiest magnates. The title of earl could even be used for the ruler of the whole country, as in the case of Earl Hákon. For the most part, however, Snorri distinguishes relatively clearly between the person or persons claiming the lordship over the whole country and the rest of the aristocracy. His definition of the aristocracy, however, is more problematic.

The Aristocracy

Snorri gives several titles for members of the aristocracy, both before and after Haraldr hárfagri's conquest of the whole country. Having described this conquest, he gives a brief account of Haraldr's system of government: Haraldr put one earl in each district called fylki as its supreme judge, who was to have one-third of the royal incomes in his district. Under him there were to be four magnates of lower rank (hersar ) with fixed veitslur —twenty marks—in income. Both the earls and the hersar were required to muster a certain number of men for the king's army. According to this description, the aristocracy consisted of royal officials, organized as a hierarchy (HHárf . chap. 6). Later he tells that this arrangement was upheld for a long time, and that Earl Hákon of Lade had sixteen earls under him (OT chap. 45—with reference to a scaldic stanza). Under Haraldr harðráði he refers to a custom, introduced by St. Óláfr and Magnús góði, of having only one earl at a time in the country (HHarð . chap. 48). Though this alleged rule may be introduced for the sake of the story Snorri was just then telling, it conforms better with Snorri's account of actual practice than the system of one earl in each fylki , as he only mentions a small number of earls at any one period (see Bøe, 1962: 560 f.). Nor does the number of hersar in Snorri's story actually conform to the enormous number that must have existed if Haraldr's rule had been practiced. To Snorri and other Icelandic saga writers, the term hersir seems to have been equivalent to lendr maðr , which is the usual term from the saga of St. Óláfr on (Bøe, 1965: 498 ff.). Snorri implies that these positions were not hereditary, although custom favored certain kins.

The real aristocracy is not confined to men with these titles. In his narrative, Snorri repreatedly refers to men as esteemed, mighty, and so forth, who have a strong position in their regions and act as spokesmen for the people and local leaders. Such men may or may not be attached to the king's service as lendir menn . The term "magnate," which I have used frequently above, is meant to include these men as well as those holding formal titles. In this way, the aristocracy is vaguely separated from the king on the one side and the people on the other. Though titles are not without importance, they are clearly insufficient to distinguish Snorri's aristocracy from the rest of the population. We may note, however, that in contrast to the office of goði in the society to which Snorri himself belonged, medieval Iceland, there was no formal office for local leaders, except those distributed by the king.

Despite the rather vague lines of distinction between the categories, there is no real doubt of the tripartite structure in itself. As is evident both from the discussion above and from Sandvik's analysis, the magnates are the clue to the whole system. They are the intermediate category, with connections to the king above them and the people below. Sandvik emphasizes their function as leaders of the people in opposition to the king, in my opinion too strongly. As we have seen, his and Koht's picture of an almost constant conflict between the latter and the people is hardly correct. Still, the contrast between the magnates' almost completely harmonious relationship to the people and their frequent conflicts with the king is striking and demands an explanation. Before we turn to that, however, we shall have a closer look at the foundation of their power.

We can distinguish between four elements as the foundation of a magnate's power: (1) kin; (2) wealth; (3) charisma; (4) office bestowed by the king. All four can be illustrated in the person of Erlingr Skjálgsson. Erlingr was a descendant of Hordakari^[*] , who was an important man in Western Norway in the mid-tenth century, and whose kinswoman Þóra became Haraldr hárfagri's mistress and the mother of King Hákon góði. His son, Klyppr hersir, killed King Sigurðr Eiríksson for having raped his wife (HGráf . chap. 14). Snorri gives Erlingr's genealogy in connection with Óláfr Tryggvason's visit to Western Norway to convert this region.  Hordakari's kin is presented as the mightiest there (OT chap. 54). Under their leader  , they agree in accepting Christianity on the condition that Erlingr is married to Óláfr's sister, a marriage that makes Erlingr the most prominent man of the region.

Erlingr is thus elected to succeed   as the leader of the clan. He is the grandson of Hordakari's second son, whereas   is Hordakari's fourth son, probably the only one surviving at this time. His grandson, Áslákr fitjaskalli, later becomes Erlingr's enemy and kills

him. Theoretically, this might indicate some sort of primogeniture within each generation of the clan. But there is no trace of such principles in Snorri's own Icelandic society (see n. 16), nor does Snorri refer to it in this case. He clearly states that Erlingr is elected because of personal qualities, that is, charisma .

Snorri normally seems to imply that the magnates have inherited their position, and he often gives their genealogies, usually when introducing them. Sometimes he does so in order to trace the descent of men who have played an important part recently or in his own time, such as Archbishop Eysteinn and his own friend Duke Skúli (HHarð . chaps. 37, 98: 219). Modern scholars, who have tried to establish genealogies on the basis of Snorri and other historians, have succeeded in linking together a number of leading magnates from various epochs, which gives the general impression that there was a considerable continuity in the aristocracy from the time of St. (Óláfr until the latter half of the twelfth century.^[22] The self-esteem of the old aristocracy is evident from the reaction of Hárekr at Tjøtta against King Óláfr's promotion of Ásmundr Grankellsson, which in Hárekr's opinion means reducing the status of the old aristocracy in favor of a "new man."^[23] In the same way, the position of the low-born Selþórir is an extreme provocation of Erlingr Skjálgsson. But Snorri does not always point out the links that must have existed, which indicates that he is less interested in genealogy than in other criteria of aristocratic status.

There is some connection between charisma and descent: noble blood will normally produce a noble character. The idea is directly expressed by Earl Rognvaldr^[*] of Møre, who says to his son Einarr that he will hardly bring any honor to his kinsmen, because his mother was a slave.^[24] In this case, Rognvaldr is actually proved wrong: Einarr defeats the Vikings who have driven his half-brother away from the Orkneys, establishes himself as earl there, and becomes a mighty ruler. His words do not therefore necessarily express Snorri's own opinion. Snorri even gives an example of a slave becoming king by rebelling and deposing the right king.^[25] However, Snorri's characterizations are best interpreted in the light of some belief in a general connection between noble blood and noble character, though Snorri does not necessarily mean that this always holds true: this idea is balanced by a rather matter of fact consideration concerning courage: younger and less established people with more to gain are likely to be braver than the great magnates (OH, chap. 90: 167).

With the relatively vague distinction between the magnates and the rest of the population, what kind of blood is considered noble? When Snorri refers to a man as gofugr^[*] (great, prominent) or ættstórr (highborn) he probably refers to the fact that his ancestors have been wealthy and mighty men in a particular region over some generations. One

qualification, however, seems to be particularly important, namely connection with the royal family. In his remark on the great power of the lendir menn at the beginning of St. Óláfr's reign, he explicitly mentions that many of them were descended from kings or earls (OH chap. 46). This is borne out by individual examples of men who are related to the king or the earls of Lade through kinship or marriage, such as Erlingr Skjálgsson, Einarr þambarskelfir, and Hárekr at Tjøtta. When the latter complains about Óláfr's promotion of the "new man" Ásmundr Grankellsson, he may very well have this in mind. Ásmundr's father is characterized as an auðigr bóandi (a wealthy farmer), a former Viking chieftain and a hermaðr mikill (a great warrior) (OH chap. 106), whereas Hárekr is himself something of a newcomer as far as wealth is concerned (OH chap. 104). The great difference between the two is apparently that Hárekr is related to the royal family, whereas Ásmundr is not. Membership in the royal family is in itself a sign of prominence, whereas families who rise to prominence through wealth and their own efforts will sooner or later end up with being married into it.

Snorri repeatedly refers to Erlingr's great wealth. Erlingr is sufficiently wealthy to be the leading man in Western Norway even without the king's veitslur (OH chap. 22). His wealth evidently consists in large estates, to which Snorri refers in the story of his allowing his slaves to buy themselves free through extra working (OH chap. 23; see also chap. 117). In one sense this is a further indication of the importance of birth for aristocratic status, as landed estates were hereditary. But Norwegian and to some extent Icelandic rules of divided inheritance made it less likely than in systems of primogeniture that the position of a particular family or kin would be maintained (see n. 15). Besides, Snorri also mentions other methods of acquiring wealth, above all Viking expeditions. Though one probably had to hold a prominent position within one's region to become the leader of such an expedition, its success or failure very much depended on luck and personal ability. Such expeditions must therefore have contributed to some degree of social mobility. The civil wars of the late twelfth and early thirteenth centuries may have had the same effect.^[26] Even landed estates might be acquired through personal efforts, as the example of Hárekr at Tjøtta shows.^[27] In Erlingr Skjálgsson's case, Viking expeditions compensated for his loss of royal veitslur after the death of Óláfr Tryggvason. Politically, wealth served two purposes: to keep a numerous and well-equipped army of retainers and to attract adherents and allies through generosity.

The importance of charisma appears from Snorri's description of Erlingr, which makes him a sort of popular hero among Snorri's readers even today. His old kinsman   calls him the most beautiful young man in Norway and Óláfr Tryggvason points out his beauty to his reluc-

tant sister (OT chap. 56). Snorri repeatedly stresses his skill at arms, bravery, and attractive personality, which makes him a number of friends. His last battle is one of the great "scenes" of Heimskringla , with Erlingr standing alone on his ship after all his men have been killed, defending himself "so nobly that no one could remember any one man having stood off the attack of so many so long" ("svá prúðliga, at engi maðr vissi doemi, at einn maðr hefði staðit svá lengi fyrir jammargra manna atsókn") (OH chap. 176: 405). In his narrative, Snorri often implies that a magnate's ability to win adherents and thus to improve his position in society depends on qualities like eloquence, wisdom, bravery, and so forth.

Erlingr's position under the earls and Óláfr serves as an example of a magnate managing to be extremely powerful even without the king's support. This should not, however, be taken as evidence that the king is normally of marginal importance for the position of the magnates or that there is usually a conflict of interest between them. Snorri clearly implies that Erlingr is an exception in this respect. His counterpart in Trøndeiag, Einarr þambarskelfir, is apparently without veitslur after his reconciliation with Óláfr. He is still a very rich man but plays no political role until after Óláfr's exile (OH, chaps. 121, 144, 171, 194). In other cases, such as in most of Óláfr's alliances with "new men," the king's support is decisive in promoting people to high rank in society, though it is clearly not a question of raising men from the dust. Snorri even expresses this point in a more general way: Haraldr hárfagri's conquest and his new arrangement of government gave his "officials" more wealth and power than the petty kings had had before, thus leading many magnates to Haraldr's service. Even Erlingr is to a considerable extent an example of a man promoted by the king. During the negotiations between Óláfr Tryggvason and Erlingr's kinsmen about his marriage with Óláfr's sister Ástríðr Ástríðr, first refuses to marry Erlingr because he is an útíginn maðr , that is, a man without aristocratic rank. Óláfr offers to make him earl, but Erlingr refuses, not wanting a higher name than that of his kinsmen before him, who have been hersar . But he wants Óláfr to make him the mightiest hersir in the country, which Óláfr in fact does. Erlingr in reality becomes the prince of most of Western Norway and is able to maintain this position against Óláfr's successors. Even great magnates, like Þórir hundr and Hárekr at Tjøtta, to some extent owe their position to the king.^[28] Another interesting example is Kálfr Árnason, whom Óláfr gives the estate of the former local leader Olvi^[*] at Egge and who has apparently no difficulty in maintaining his position as the leading magnate of Inner Trøndelag.

It is difficult to form an exact opinion on the relative importance of Snorri's various criteria for aristocratic status. Snorri does not treat the question systematically, nor is he sufficiently specific in his detailed nar-

rative to allow us to weigh their importance, for instance the size of income derived from royal veitslur versus "private property," the importance of hereditary position versus "appointment" by the king or the people. The previous examples are, however, sufficient to reject Sandvik's view of the magnates as exclusively popular leaders. Snorri implies that they derive a considerable part of their power from the king: to be appointed lendr maðr by the king means a clear promotion compared to other men of equal wealth and status and even gives considerable economic resources. And membership in the royal family by kinship or marriage is the highest expression of aristocratic status. Snorri thus regards the magnates as mediators between the king and the people rather than exclusively popular leaders.

The King

The ideology of the Norwegian monarchy of the thirteenth century, as expressed in the official sources, notably The King's Mirror , was derived from the European and Christian idea of the king as God's representative on earth: the king holds his office from God. He has duties to perform on God's behalf, above all in the field of justice. He is given extensive power by God in order to perform those duties, the most important check on him being that he is not allowed to reduce his power. And finally, he rules a hierarchically organized society, whose members owe him loyalty and obedience, the more so the higher their rank (Bagge, 1987b : 22 ff., 50 ff., 153 ff., 205 ff.).

As might be expected from my previous analysis of the conflicts, Snorri's picture of the king differs from this ideology on most points. As for legitimacy, both the traditional Christian idea of the king as God's representative on earth and the more recent idea of the king as the representative of the state, are largely absent from Snorri's work. Although the latter is not very surprising, the former idea is found in Norwegian sources long before Snorri's time and is clearly expressed by the clerical historians of the late twelfth century.^[29] Snorri's picture of St. Óláfr as the rex iustus belongs to this tradition but, as we have seen, it is of minor importance in Snorri's overall view of his reign.

The supernatural aspect of kingship in Snorri is rather an example of what Max Weber calls charismatic lordship ("charismatische Herrschaft"), which includes both particular powers inherent in the royal blood and impressive personal qualities (Weber, 1964, I: 179 ff.). Though Snorri most probably believed in supernatural qualities inherent in the king, especially the king's luck, his main emphasis is on charisma in the sense of ordinary, human qualities that make the king attractive to other men.

The idea that the king should belong to the royal line of descent may possibly be derived from the belief in particular powers inherent in the royal blood. Sigurðr slembir, who was the most able of the pretenders in the first phase of the civil wars, had serious difficulties when he could not persuade people of his royal blood. Even Magnús Erlingsson's claims on the throne were allegedly dubious because he only descended from the royal line through his mother (ME chap. 21: 462). However, royal descent apparently plays no role for the magnates who rebel against King Magnús berfoetr under the leadership of Sveinn Haraldsson, a man of Danish noble descent, whose genealogy Snorri does not attempt to trace (MB chap. 4).^[30] Nor does Snorri reflect on the legitimacy of Earl Hákon Sigurðarson's ascent to rulership over Norway after the fall of King Haraldr gráfeldr or Haraldr hárðráði's claim while his nephew, King Magnús, was still living.^[31]

Snorri's inconsistencies at this point probably reflect actual historical development. Royal descent became progressively more important and the way the kingdom was transmitted from one generation to the other was more clearly defined.^[32] Snorri and other thirteenth-century sagawriters seem generally to express the stricter attitude of their period, and there are reasons to believe that some of their royal genealogies of the early period are actually constructions. Occasionally, however, the realities of the earlier age appear in their narrative. Thus, although it is possible though not usual for "new men" to achieve aristocratic status and relatively easy to succeed a local leader without being blood related to him, the qualifications for kingship seem to be stricter.

In the official ideology of the Norwegian monarchy, God's election of the king is expressed through hereditary succession (Bagge, 1987b : 39f.). This may even suggest that the idea of hereditary succession is not derived from the traditional belief in the powers inherent in the royal blood, but is of fairly recent and Christian origin. This is not a necessary conclusion, however. What is characteristic of thirteenth-century ideology is not hereditary succession as such but its restriction to one person in each generation and the very definite rules for selecting this person. This idea is in turn combined with the legal or quasi-legal idea of the kingdom as the hereditary property (óðal), of the royal line (Bagge, 1987b : 31 ff.). By contrast, Snorri seems to believe in particular powers inherent in the royal blood, at least in the sense that royal descent is more likely to produce the qualities needed in a king than descent from ordinary men. But such powers evidently belong to all the king's descendants, not only to one heir. Moreover, they are probably not qualitatively different from the ones belonging to other great men. A man's descent is important for his personal qualities. The king ought to be the best man in the country, and the chance for this is better if he belongs to the royal line.

There is little trace in Snorri of the ecclesiastical and monarchical idea of the king holding an office on God's behalf (see Bagge, 1987b : 25 f., 194 ff,). Snorri does not refer to the king's position as an office (embætti etc.), nor does he distinguish between the king's person and his position within society. His position is to be regarded as a social role rather than as an office in the modern or contemporary ecclesiastical sense: to be king is first of all a personal performance, and it is difficult to distinguish between the royal person and the royal office. However, his position is an office in the sense that he is formally appointed to it, through election at the popular assemblies; that there is normally only one holder of it at a time—if there are more, this is a problem that needs a special solution—and that certain rights and duties appertain to it. In this way, the king's position differs from that of the magnates. Though most magnates are also formally appointed to an office, that of lendr maðr , this is not as essential to their position as the king's ascendance to the throne.

The idea that the king has his legitimacy from the people is expressed in his formal election at the popular assemblies which is frequently mentioned in the sagas. This popular right of election was progressively restricted during the high Middle Ages and had degenerated into acclamation by the latter half of the thirteenth century (Helle, 1974: 60 ff., 115 ff., 136 f.; Bagge, 1987b : 40; see also NG chaps. 4-8, L part 2, chaps. 3-6, H chaps. 2-10). Snorri seems in this respect to reflect the more "democratic" attitude of the earlier Middle Ages, usually mentioning with great care the kings' acclamations at each particular assembly. As nearly all pretenders are acclaimed at least at some assemblies, however, the real decision between rival pretenders is not taken there, but in more informal encounters between them and their adherents, and ultimately at the battlefield. However, this is also a kind of "democratic process," as success in such encounters depends on the ability to attract adherents. Thus, although Snorri tones down the mystical and magical aspect of charisma and its hereditary character (in Weber's terminology, "das Erbcharisma"; see 1964: 183 f.), he lays great emphasis on its personal aspect. The king is supposed to be an impressive personality. As rival pretenders are all of royal blood, and it is impossible on the basis of heredity alone to decide which one of them has the best right, personal qualities are usually decisive. Wealth has a similar importance. Gifts are necessary to attract adherents, and the king must therefore be both wealthy and willing to spend his wealth.

In the version of the Christian ideology of the rex iustus that was shaped by the thirteenth-century Norwegian monarchy, jurisdiction is the single most important duty of the king; in The King's Mirror it even seems to be his only duty (Bagge, 1987b : 51 ff., 156 ff., 189 ff.). Through its insistence on the king's position as the supreme judge of the realm,

The King's Mirror also lays the foundation of a theory of his legislative authority, though this is first explicitly stated in the laws of the second half of the century (Bagge, 1987b : 156 ff., 174 n. 28, 189 ff.; see e.g., L Prologue, part 2, chap. 8).

This picture of the king is occasionally found in Snorri. In the competition between the kings Sigurðr and Eysteinn, a fairly long passage is devoted to the performance of the two kings in the field of justice and administration (below, chap. 4). Snorri's picture of St. Óláfr's daily life as king resembles that of the rex iustus in The King's Mirror : he rises early, attends mass and the canonical hours, and spends his day in meetings and in settling peoples' issues (OH chap. 58; see also Kgs .: 92, lines 8-34, 97.12-98.19 and Bagge, 1987b : 90 ff.). He is also a great legislator, in accordance with contemporary tradition, which attributed much of the existing legislation to him (see Blom, 1981: 61 ff. with ref.). Further, Snorri often refers to the king's duties as a judge and legislator in his characterizations of other kings. King Hálfdan svarti was a just man who issued laws and respected them himself and who stipulated the fines for various offenses according to the rank of those involved (HS chap. 7: 95). He was thus apparently the inventor of the system of fines that still existed at Snorri's time. King Haraldr hárfagri made arrangements for the governance of the country, including the judicial administration (HHárf . chap. 6), but apart from this, nothing much is said of his judicial activity. By contrast Hákon góði was the great legislator, who issued the   and   (HG chap. 11) and organized the leiðangr (the popular levy) (HG chap. 20). After his "conversion" (Óláfr's son and successor Magnús becomes a wise king and a great legislator, the author of Grágás , which was still the law of Trøndelag at Snorri's time (MG chap. 16). Haraldr harðráði is above all a warrior king but is also mentioned as a strong ruler within the country (HHarð . chap. 36). Among later rulers Snorri praises Erlingr skakki for his strong rule, which probably includes judicial activity, and for his effort in maintaining internal peace (ME chap. 37; see also HHerð . chap. 9; ME chap. 10).

In practice, however, we rarely meet the king in what in thirteenth-century ideology appears as his normal capacity as judge and internal administrator. This is natural insofar as the sagas are interested in extraordinary events, not in normal routines. Therefore, the king's function as war leader is more prominent than his function as judge, administrator, and legislator. This applies not only to Snorri but also to the sagas dealing with the period after 1177, such as Sverris saga and Hákonar saga .^[33] What is somewhat more surprising is the fact that so few of the conflicts Snorri deals with arise out of such activities.^[34] By contrast, in the Icelandic sagas we are constantly reminded of the chieftains' position

as great farmers and local leaders and mediators through the kinds of conflicts in which they are engaged (Byock, 1982: 39 ff., and 1988: 113 ff., 168 ff.). As we have seen, the king in Heimskringla is mainly a great magnate, who tries to extend his sphere of influence by outmaneuvering his opponents and gaining adherents through generosity, patronage, and so forth. Though Snorri no doubt takes into account the official thirteenth-century picture of the king as a judge and administrator and the representative of an impersonal justice, it is not really integrated in his narrative.

As for the king's government, Snorri's description of Haraldr hár-fagri's appointment of earls and hersar in each district after his conquest of the country seems to imply some kind of bureaucratic structure. However, the importance of this should not be overrated. First, the system appears more like a regulation of the traditional system of government than as an entirely new one, and the new officials seem largely to be recruited from the old class of magnates and to govern much in the same way. Second, Snorri rarely refers to the system after Haraldr's own time.

However, Snorri frequently refers to the king's local representatives, particularly from the period of St. Óláfr onward. There were two kinds of such men, the mighty and high-born lendir menn , who seem to be the king's allies more than his officials, and the ármenn , who were of lower rank, though not necessarily descended from slaves, like Selþórir.^[35] The ármenn administered the king's farms but also had "public" duties in the district, such as jurisdiction, as appears from the examples of Selþórir and Bjorn^[*] . A lendr maðr was assigned a particular part of the royal demesne as veitsla for life, an arrangement resembling the European fief. In addition, he performed duties on the king's behalf, similar to those of the ármaðr . Snorri sometimes implies that the lendir menn had regular districts, called lén or sýslur (e.g., OH chaps. 116, 123, 167; see also Hertzberg, 1893: 307 ff., and Lie, 1907: 22 ff.). This is an administrative system similar to the one found in Norway in Snorri's own lifetime, which had a quasi-bureaucratic character and in which the officials were clearly subordinated to the king and served as his representatives.^[36] However, we rarely come across this system in practice in Snorri's narrative. There the lendir menn are mainly the king's local allies, whereas the ármenn are his personal servants.

Snorri's attitude to bureaucracy is well illustrated through the conflict between St. Óláfr and Erlingr Skjálgsson over Selþórir, Óláfr's ármaðr at Avaldsnes, that is, within Erlingr's "kingdom" of Western Norway. In one of their encounters Erlingr expresses his willingness to obey the king, but flatly refuses to submit to the low-born Þórir.^[37] Selþórir is not described as an attractive person, and Erlingr probably

has Snorri's full sympathy. To some extent, this controversy can be regarded as a conflict between the old system of personal loyalty between the king and the magnates who were his local allies and a bureaucratic system of local officials. But this aspect should not be overrated. All magnates had low-born men in their service, whose prestige reflected that of their master. The mightier the magnate, the more low-born creature could he afford to use against his rivals. Characteristically, when Þórir is killed, Óláfr is not concerned with the fact that he is a royal official, but with the offense against his own honor, because the killing took place in his presence.^[38] A parallel case is that of Einarr þambarskelfir, who risks a conflict with King Haraldr harðráði to save a thief—the most despicable of all men—from execution, because the man had been in his service. Moreover, Snorri evidently does not regard Selþórir as representative of Óláfr's normal way of building up his power at the local level. Óláfr did try to curb the power of the great magnates, but his usual method was to use men of approximately the same rank or slightly lower as their rivals.

Rather than building up a bureaucratic or quasi-bureaucratic apparatus, Snorri's kings rule by establishing personal connections, through gifts, concessions, and—not the least—through marriage, sexual relationships, or fostering. Haraldr hárfagri marries most of his daughters to his earls (HHárf . chap. 42) and lets his sons grow up with their mothers' kin (HHárf . chap. 21). In accordance with his policy of promoting second-rank men, St. Óláfr marries his female relatives to such men instead of trying to win the greatest magnates over to his cause in this way.^[39] His half-brother, Harald harðráði, coming from abroad and apparently lacking a strong kindred—and female relatives?—goes to the fairly unusual step of marrying the Norwegian magnate Þorbergr Árnason's daughter Þóra (HHarð . chap. 33). During the civil wars, the most prominent members of the factions are normally related to the royal family through kinship, marriage, or fostering.^[40] Given the importance of such relations for aristocratic status, this is probably an important instrument for the king in linking the magnates to his service. Another such instrument is membership in the king's hirð , a community that bears some resemblance to a kindred or a family: the members share the same house and table, are obliged to keep peace between them and to avenge one another.^[41] When one member of a kindred is included in the hirð , the whole kindred or at least some of its members may be drawn into the king's sphere of influence. The story of Grankell and his son Ásmundr is an example of this.^[42]

The importance of such personal links between the king and his servants clearly indicates that the people as a whole have few obligations to the king. Admittedly, Snorri agrees with royalist ideology in not defining

any constitutional checks on the king's power. Although in the latter this is combined with rather explicit statements of the king's competence and the respect and obedience the people owe to him (Bagge, 1987b : 22 ff., 156 ff.), there are few such statements in Snorri. The idea of treason against the king is expressed in Óláfr's words to Erlingr Skjálgsson after his surrender in the last fight between them: "A mark he shall bear, the betrayer of his king ("merkja skal dróttinsvikann," OH chap. 176: 406). In a similar way Finnr Árnason warns Hákon Ívarsson that he will be called a traitor if he rebels against the king and defeats him (HHarð . chap. 47). Óláfr's severity against offenders against him may also be intended as an illustration of the idea of the king as representing impersonal justice and therefore refusing to accept indemnity for men like Asbjorn^[*] selsbani and Þórir Olvisson^[*] . Though Snorri clearly knows these ideas and occasionally refers to them, they are of minor importance in his actual narrative. Usually, the king is an actor in the political game like everyone else, and Snorri rarely expects the other actors to take into consideration his exalted position as God's representative on earth. Although the king's men are obliged to serve their lord and avenge him if he is killed, the people as a whole have few such obligations. When one king is defeated and killed, they usually acclaim the victor as his successor—their refusal to acclaim Sigurðr slembir is exceptional.^[43] To get support, the king must do two things. He must rule in such a way as to promote the interests of the people, or at least not counteract them. As we have seen, Snorri gives several examples of kings being deposed or killed for not respecting the rights and interests of the people. And he must form links to the leading magnates—make them his clients—so that their clients in turn will support him.

There is thus very little qualitative difference between a king and a magnate. The king is the magnates' superior, but he depends on their support in the same way as they depend on the support of the people. Although there are no constitutional checks on him, he cannot claim special protection from rebellion and opposition by appealing to God's election. Snorri most probably believes in particular powers inherent in the royal blood, but such powers are not qualitatively different from those of the magnates. The king is a primus inter pares who depends above all on his personal ability to be able to rule. The qualities demanded of him differ more in degree than in kind from those demanded from the magnates, as is evident from the characterizations. There are traces of the contemporary monarchical ideology in Snorri's reference to the king's duties as a judge and legislator, but generally, he performs much the same functions as the magnates. Rather than a bureaucratic apparatus, he must use a net of personal alliances with more or less prominent magnates to be able to rule.

The Magnates Between the King and the People

By and large then, there is a close similarity between the king and the magnates. Both have largely the same foundation of their power, both have to rule by gaining followers through gifts and personal attraction, and the magnates do not regard the king as qualitatively different from themselves.

Admittedly, Sandvik is correct in pointing out the contrast between the magnates' frequent conflicts with the king on the one hand and their undisputed position as popular leaders on the other. There is, however, another explanation to this, namely that Snorri is primarily interested in kings—in fact, Heimskringla is strictly speaking a series of royal biographies, not what we would call a national history. Consequently, conflicts below the royal level are uninteresting. This suggestion receives some support from Snorri's casual references to conflicts between magnates that did not directly involve the king, as the one between Hárekr of Tjøtta's and Ásmundr Grankellsson's kinsmen (MG chap. 12). The background to the story of Asbjorn^[*] selsbani points in the same direction. Asbjorn's generosity has a political background. He tries to win adherents in order to become the leading man in the region in competition with his uncle Þórir hundr (OH chap. 117). Snorri evidently does not regard local society as a society of perfect stability under the leadership of magnates with a fixed position. Rather, he implies that there is rivalry and competition between the leaders there in the same way as in the country as a whole. Even a "constitutional" conflict between a magnate and the people does not seem entirely out of question. Besides being the ruler of the whole of Norway, Earl Hákon of Lade was also the local leader of Trøndelag and derived much of his strength from this position. But it was the people of this region that eventually deposed and killed him. The reason for this, his behavior toward their women, is quite the sort of oppression one might expect from a local leader who grows too self-confident.

The contrast between the relationship between the king and the magnates on the one hand and the magnates and the people on the other is therefore of less importance than appears at first sight. Nor are there sufficient reasons to believe in a constant conflict between the king and the magnates. The magnates should rather be regarded as a group between the king and the people, depending on support from both and having an important function for both. Admittedly, it seems likely that a magnate may control his people more easily within his smaller region than the king who tries to establish his dominion over the whole country and that Snorri had this in mind when telling his stories. We need not

necessarily imagine the same loose organization on the local level as in society as a whole. Kinship, personal acquaintance, and personal loyalty may account for more there and thus create more stable relationships. The fact that men of lower status are supposed to show greater loyalty to their superiors than the great magantes also point in this direction.

The difference is, however, rather one of degreee than of kind. Basically, both the king and the magnates have the same foundation of their power: they have some hereditary claims, but these are insufficient without popular support, which is obtained through generosity, success, and charisma. Society is loosely organized and support is easily lost or won. No strong organization for control over the country exists, and military means are clearly insufficient for a real conquest. Military victory is a means to win support, not an end in itself. This is the lesson that can be drawn from Óláfr's and Cnut's conquests of the country, from Asbjorn^[*] selsbani's desperate attempts to maintain generosity, and from the dramatic clashes between the king and the people at the numerous popular assemblies.

An Aristocratic Society?

From a modern point of view, it is not difficult to agree in Sandvik's description of Snorri's society as an aristocratic one. It is commonplace in Heimskringla that the people are helpless without their leaders. When the king manages to kill the local leader in battle or through a coup, popular resistance is impossible. Even if they are numerically superior, the people are unable to avenge their leader, as the case of Einarr þambarskelfir's fall and numerous other examples show (HHarð . chap. 44; see also Sandvik, 1955: 49 ff.). During the civil wars, the great magnates are the most able leaders, and the faction that has the support of the majority of them is most likely to win. At the popular assemblies, there is always a magnate to speak for the people, and the magnates' opinions are decisive for the final outcome of the deliberatons (Sandvik, 1955: 45 ff.). When St. Óláfr offers the Icelanders his friendship in return for Grímsey, an uninhabited island off the coast of the country, Guðmundr Eyjólfsson at   almost persuades the assembly to grant the king his prayer, until his brother Einarr stands up and speaks against the demand in a way that turns the opinion entirely against the king (OH chap. 125; see also Sandvik, 1955: 33 f.).

The opposition against the Danish rule after the death of St. Óláfr is in one sense an exception to this (OH chaps. 247, 250-251; MG chaps. 1-5). The Danes are very unpopular, and the widespread belief in Óláfr's sainthood serves to strengthen the opposition. After the young King Magnús's arrival in Trøndelag, King Sveinn tries to mobilize the

people of Western Norway at an assembly in Sunnhordland. But few men appear, and King Sveinn and his Danish adherents receive little acclamation at their speeches and some even speak against them. In a short speech, Sveinn draws the conclusion that those who are not present or have not spoken in favor of Magnús will hardly be more trustworthy than those who have and decides to return to Denmark. Snorri tells the episode rather briefly, neither naming the leaders of the opposition nor stating their arguments in direct speech. This may indicate that they were not magnates in the real sense, though they may well have been prominent men at the local level. The fact that he refers to them as boendr (commoners) is less significant from this point of view. Though this is hardly an example of a spontaneous, popular uprising, it shows local society in opposition against the ruler. But Snorri seems to imply that it is rather exceptional that ordinary men dare to speak against the king in his own presence.

Taken as a whole, however, Snorri's account of the rebellion against the Danes clearly emphasizes the role of the two magnates Einarr þambarskelfir and Kálfr Árnason more so than those of his predecessors. Einarr, who is disappointed at not being rewarded as he was promised, begins early to exploit the rumors of Óláfr's sainthood. A meeting between him and Bishop Grimkell results in Óláfr's canonization. Later he allies himself with Kálfr Árnason, and the two of them become the leaders of the opposition and later bring Óláfr's son Magnús back from Russia and make him king.

On the explicit level, Snorri attaches great importance to the aristocracy. When we read his text as a whole, however, the importance of the people is equally striking. This is above all evident through a comparison with contemporary European sources. In European historiography of the period the people are hardly mentioned at all, whereas in Heimskringla and other sagas they are constantly present. Against a contemporary European background, Snorri's constant reference to the popular assemblies, both concerning royal elections and other matters, is more remarkable than the fact that the people at the assemblies are normally led by their superiors. The difference becomes even more striking when we consider the interest aspect of the policy of the kings and magnates.

One of the reasons why Snorri implies that there is normally no conflict between the magnates and the people is that the magnates become leaders largely by serving the interest of the people. The magnates who stand up against the king at the assemblies, like Asbjorn^[*] of Meðalhús, Þorgnýr  , Einarr Eyjólfsson, and Sigvatr skáld, articulate common interests. The kind of interests that is most frequently in focus is peace. Snorri generally assumes that the people want peace, and thus peaceful kings are more popular than those who conduct an

aggressive foreign policy, like Haraldr harðráði and Magnús berfoetr (HHarð . chaps. 36, 99; MB chap. 26; see also above). A magnate like Erlingr skakki, who is the real ruler of the country in the 1160s and 1170s, is popular with the people because of his ability to create internal peace through suppressing the raiding groups of the opposite party. Peace negotiations are very often the result of initiative from the people. This applies to the conflicts between St. Óláfr and King Óláfr of Sweden and between Earl Hákon of Lade and the Eiríkssons (OH chap. 68; HGráf . chap. 6). In these cases the magnates are often the people's spokesmen before the king.^[44] Second to peace, the people want good years and low taxes. They support the rulers they think are able to help them to achieve this, and, led by the magnates, they rebel against those who are not. Finally, the people are usually conservative, for instance in religious questions. Led by the magnates they oppose the kings' attempts at conversion. But when the magnates convert, they usually follow. Dealing with politics at a high level and with dramatic episodes, Snorri usually focuses on the magnates' role as leaders in the conflicts that result from these events. Occasionally, however, he alludes to the magnates' more normal function of leadership, as offering hospitality (e.g., Olvi^[*] at Egge and Asbjorn^[*] selsbani, OH chaps. 107-110, 117) and protecting their clients or adherents. His reference to what was probably a very important aspect of leadership, settling legal issues between people, is more vague, but he does mention their judicial functions in connection with their conflicts with the kings. Erlingr Skjálgsson appropriates the fines in his district whether the rulers agree or not (OH chap. 22; see also chap. 116) and Einarr þambarskelfir's conflict with King Haraldr harðráði takes place at a popular assembly and concerns legal matters (HHarð . chaps. 43-44).

Snorri thus represents an aristocratic view in the sense that he considers the common people unable to articulate their own interests, but not in the sense that their interests are unimportant. Admittedly, aristocratic, self-interest can easily be presented as the common good of society. Snorri, however, presents both popular interests in such a concrete way and so frequently refers to the people and their opinions that he clearly represents a different and more "democratic" historiography than the one current in contemporary Europe.

Snorri distinguishes between two types of kings, the warrior hero and the peaceful ruler, the latter of which is clearly the more popular with the people. To some extent, the same distinction may be made concerning the magnates. In economic terms, the king or magnate may spend his surplus in two distinct ways in order to get support, either on an army of retainers or on ordinary people. Snorri sometimes points to the generosity of kings to their hirð as opposed to their harshness toward the peo-

ple. Although he mentions magnates, such as Erlingr Skjálgsson and Grégóríús Dagsson, with large armies of retainers to whom they are very generous, the distinction seems less relevant in their case. All together Snorri's description of the road to victory clearly implies that exclusive reliance on an army of retainers is not sufficient, but is more likely to lead to disaster. Popular support is necessary, to kings as well as to magnates.

Against this background we must see the "constitutional" conflicts we have considered above. Given his view of the political game as a question of winning support, Snorri evidently regards it as very important for a king to treat the people in the right way and finds it natural that they react against hard exactions, such as Haraldr hárfagri confiscating their óðal , Sveinn and Álfífa's taxes and strict laws, and Magnús góði's harsh measures before his "conversion." To Snorri the people are both strong enough and sufficiently organized to react against royal oppression. But this does not mean that he regards such conflicts as the main theme in his history. They are rather one of several kinds of conflict, and, to judge from their relative infrequency, hardly the most important.

The Social Order

The difference between Snorri's idea of society and that of the contemporary Norwegian monarchy becomes evident through a comparison with sources like A Speech Against the Bishops (c. 1200) and The King's Mirror (c. 1255). A Speech Against the Bishops is the product of the violent conflict between King Sverrir (1177-1202) and the Church toward the end of his reign. Its author sets out to prove from "the holy writs," that is, ecclesiastical authorities contained in Gratian's Decretum , that the king should have a leading function within the Church. Although this source deals explicitly only with the Church, it contains an overall view of society which is relevant in other contexts as well.

This view is most directly set forth in the image of the human body, which represents the Church, that is, the Carolingian ecclesia, comprising both ecclesiastical and secular society (Holtsmark, 1931: 1, line 11-2, line 7; see also Gunnes, 1971: 50 ff., 73 ff.): Christ and his holy Church are one body, and the author then goes on to describe how the various limbs and organs, which represent clerical and secular ranks and offices, take care of different functions. The difference between the clergy and the laity is fairly clear, the former representing the senses and the organs of digestion and serving to give the body spiritual nourishment and guidance, whereas the latter represents the skeleton and the muscles and should defend the body. A strict hierarchy prevails within this part of the body. The king is the chest and the heart and conse-

quently the real governor. The rest of the aristocracy is ranked according to its proximity to the king. Earls and "great princes"   are the shoulders and the back, the lendir menn are the upper arms, hirðmenn , knights, and other military men are the forearms and hands, and finally, the people are the legs and feet.

Such representations of society as the human body were very popular at the time, and many others are preserved in European sources (Gunnes, 1971: 73 ff., 367 ff.; Struve, 1978: 123 ff.). However, it is not just a question of an illustration or representation, such as may occur in our time as well, but of the idea that society is a phenomenon of the same kind as the human body, that is, there is no fundamental distinction between nature and society. The image of the human body is thus an expression of a basic idea in many traditional societies of an order that manifests itself in the whole known world, that of nature as well as that of men, and to which men have to adapt themselves. Admittedly, they do not always do so, and a given society may even over a longer period depart from this order.^[45] But this is fundamentally a moral problem. The point of "constitutional" thought, as expressed for instance in the mirrors of princes of the period, is to make the various individuals and groups acknowledge their place within this order and act accordingly. If not, the body of society is ill, as is demonstrated by the author of A Speech Against the Bishops , who explains this illness by the errors of the clergy (Holtsmark, 1931: 2, lines 8-3; line 18; see also Gunnes, 1971: 83 ff.).

The medieval constitutional thought derived from the clerical tradition differs from the modern one in its idea of the constitution as belonging to an eternal order, ultimately sanctioned by God, to which men must conform. Constitutional struggles are therefore not the expression of conflicting class or group interests or ideas, based on empirical observations, as to what sort of government is most suitable to society as a whole. They are interpreted in moral terms, as a question of conforming to God's will concerning social or political arrangements. However, this by no means prevents conflicting opinions or interests regarding constitutional arrangements.

In The King's Mirror , this view of constitutional arrangements is directly used to attack the old social order—the one that is largely represented by Snorri. Using an image of famine, in which the author compares a country ridden by civil war with a farm affected by famine, he contrasts the right social order, with a strong monarchy and public justice, with the chaos, dissolution, and feuds prevailing when the kingdom is divided between several petty kings. Then rivalry arises between the kings, they try to attract adherents from one anothers' territory, including criminals, who then escape their just punishment. This in turn

leads to general disregard for law and order, people demand compensation or take revenge for kinsman who are killed for their crimes. This leads to feuds, which ultimately lead to full civil war (Kgs. : 51, line 1-55, line 30).

What the author does here, is to present the traditional social order, according to which killing and violence are matters between the parties concerned and feud is a legal way of seeking compensation, as a deviation from the right, divinely instituted order (Bagge, 1987a : 43 ff.; 1987b : 81 f., and "Science and Political Thought"). We thus have a clear example of a "constitutional" debate, expressed in the terms of an eternal order. This implies a concept of society which is in one sense closer to modern ideas in emphasizing that society is arranged in a particular way and at least to some extent regarding this arrangement as the subject of conflicts. Nevertheless, the "constitutional" thought of these authors is fundamentally different from the modern one. The right social order is not established empirically, it is the eternal order of things. Admittedly, it may be supported by empirical arguments, as is done in The King's Mirror , but this is less necessary than in modern constitutional thought.

This idea of society is to be found more implicitly in historical works such as "the Norwegian synoptics" and Fagrskinna . A characteristic example is Theodoricus Monachus's reasons for ending his work at the death of King Sigurðr jórsalafari (1130): he does not want to tell of the crimes, murders, violations of women and sacred places, and so forth that took place in the subsequent period.^[46] To Theodoricus, armed conflicts are not part of the social order; they are its direct antithesis and against the will of God, who wants pax to prevail on earth and in society. The use of arms is only permitted for a good and just cause. Although this statement by Theodoricus seems to suggest a neutral attitude to the civil wars,^[47] his basic idea could also be used for political purposes. As a matter of fact, it became prominent exactly when the civil wars were at their most fierce. First Erlingr skakki and his faction and then Sverrir maintained that they were fighting for God or the just cause and that their adversaries were evil men who opposed the right order of society. This ideology led to Erlingr skakki's condemnation of his enemies, which shocked Snorri so much.

This moralistic attitude to internal conflicts is more prominent in the Norwegian tradition than in the Icelandic one. Theodoricus and the other historians belonging to the milieu around the archiepiscopal see of Nidaros consistently classify the kings as good or evil according to the ecclesiastical schema of the rex iustus/iniquus and carefully expose the moral relevance of their acts. The author of Fagrskinna , who is in one sense equally "secular" as Snorri in avoiding nearly all references to

religious matters and supernatural phenomena, is nevertheless closer to this tradition. Though both Snorri and many other Old Norse historians have a nationalistic bias, Fagrskinna goes further in this direction, in celebrating Norwegian deeds at the cost of the country's enemies and in derogating the Danes (Indrebø, 1917: 260; Jakobsen, 1970: 101 f.), and above all in its selection of materials. Fagrskinna is about one-fourth of Heimskringla's length and contains almost as much material on foreign policy, whereas inner Norwegian conflicts are treated more briefly (see Indrebø, 1917: 257 ff.). Very few feuds are mentioned. The conflicts that are mentioned are often of a constitutional or ideological character. Though Snorri often pays exceptional attention to such conflicts whereas the author of Fagrskinna refers to them rather briefly, there are indications that the latter also takes a different attitude.

Like the other sagas, Fagrskinna classifies the kings as good or evil, but is usually more reluctant in blaming them (Indrebø, 1917: 198 f., 223 ff., 226 f., 275 ff.). In the saga of St. Óláfr, the conflict between Óláfr and the magnates, which is also there in the earlier sagas, is made the main theme. In Fagrskinna the magnates take the initiative to get Cnut's support against Óláfr, whereas their rebellion is the result of Cnut's gifts in the earlier sagas. Both in the saga of St. Óláfr and in that of Haraldr harðráði, the author of Fagrskinna refers to the great power of the magnates and their unwillingness to submit to the king as a reason for conflict. In the case of St. Óláfr, their rebellion is aimed at his strict justice. In the reign of Haraldr harðráði, his conflict with Einarr þambaskelfir, which is presented as a feud in Heimskringla , is a rebellion against the king in Fagrskinna .^[48] Thus, the author of Fagrskinna focuses on Norwegian national unity versus other countries and on the kings as representing this unity.

This attitude in Fagrskinna is largely the result of the author's connection with the Norwegian royal court (Indrebø, 1917: 275 ff.). He may be influenced by Sverrir's struggle with the majority of the Norwegian aristocracy, but probably above all by the attempts of the contemporary monarchy to suppress feuds and strengthen public justice. Still, it would be an exaggeration to see this "constitutional" struggle as the dominant theme of Fagrskinna . His portrait's of kings are by no means always favorable, nor does he consistently side with the king in internal conflicts.^[49]

Unlike modern historiography committed to particular constitutional or political ideas, the development of this social order and its conflict with other ways of organizing society are not traced through history. It is an absolute norm, from which deviations are presented in moral terms. Although a certain pro-monarchical tendency can be detected in Fagrskinna , this ideology does not necessarily force its author always to

side with the king. The fact that individual kings may occasionally lapse from their moral responsibility does not affect divinely instituted monarchy. Nor are historical arguments necessarily very effective from a propagandistic point of view.

In contrast to works like A Speech Against the Bishops, The King's Mirror , and contemporary European mirrors of princes and political treatises, the social order is rarely mentioned explicitly in Heimskringla and nothing is said about its divine origin. This evidently does not mean that there was no such order, but rather that it was taken for granted. The apparently war- and faction-ridden society was held together, not only by a common nationalistic sentiment directed at outside attacks or by occasional explicit or implicit rules of "fair play," but by a social order that was generally accepted and not subject to conflict. There was a social hierarchy, not very strictly defined and with some mobility upward and downward, but with generally accepted rules as to who should lead and who should obey and who were allowed to fight for supreme power in society. However, this model of society gives no room for a bureaucracy, whereas the royalist model analyzed above is easily compatible with it and may even be said to include the idea of such an organization. And while the king in Heimskringla is superior in rank to the magnates, his actual power is vaguely defined and to a large extent depends on his own personality. The leaders have the right to rule by virtue of their descent, wealth, and personal qualities but they are also supposed to rule in the interest of their inferiors. As they depend on the latter for support, they are easily deposed if they do not. Thus the conflicts take place within a fairly stable, harmonious, and "democratic" society. The basic features of the social order are not contested, and victory or defeat in the conflicts between the rulers does not lead to great changes. The really important aspects of the social order are therefore left implicit in Snorri's narrative.

Finally and most important: the basic idea of society in Heimskringla corresponds fairly well to the picture in The King's Mirror of a society afflicted by famine. Feuds and revenge are taken for granted, and a man is largely valued according to his ability in conducting such conflicts. Snorri's view of society is thus in clear contrast to the royalist one. His censure of Erlingr skakki's condemnation of his adversaries may even appear like its direct refutation. Nevertheless, Snorri also shows traces of the opposite view, notably the picture of St. Óláfr as the rex iustus and the explicit explanation of his fall as the result of his strict justice.

It is not quite easy to grasp the relationship between these two views of society in Snorri. It is hardly a question of explicit royalist ideology versus traditional ideas of society as an implicit or unconscious framework. The view of society which I have traced in this chapter may be

implicit in the sense that Snorri does not regard it as an ideology as opposed to other ideologies, but hardly in the sense that he is unable to draw explicit consequences from it. Quite the contrary, Snorri uses it systematically to analyze human behavior and manages to create a largely consistent picture of politics and society in older Norwegian history. His explicit references to the rex iustus -ideology and the hierarchically organized society propagated by the contemporary Norwegian monarchy are rather to be explained as respect for King Óláfr's reputation as a saint or as concessions to Snorri's Norwegian public. Evidently, Snorri has not been fully aware of the contrast between the two ideologies. This may seem strange to us, but is more understandable in a society and political milieu in which one was not accustomed to thinking in different political ideologies. However, he may have been well aware of particular consequences of the royalist ideology which he found unacceptable, such as Erlingr skakki's attempt to use religion to crush his adversaries or arrogant and authoritarian behavior from the kings, which would inevitably provoke rebellion, and may partly have intended his work as a warning against this and more generally as an education for a politician in the fairly "democratic" and loosely organized society with which he was familiar.