Introduction
The Problem
This book deals with Heimskringla , the most famous of the Old Norse collections of kings' sagas, written by the Icelander Snorri Sturluson in the first half of the thirteenth century, probably around 1230. My intention is to examine the work's picture of man, society, and politics as well as Snorri's approach to his subject matter against a European background. The purpose of this examination is partly to characterize Snorri as a historian, partly to analyze ideas about society and politics in early thirteenth-century Norway and Iceland, and to some extent even to contribute to our knowledge of Snorri's society as it actually was.
Since the last century, in connection with the national movement after the separation from Denmark in 1814, Heimskringla has held a special place in Norwegian culture. It was for a long time one of the most widely read books in Norway and is still regarded as a kind of national symbol. Until recently, the teaching of Norwegian medieval history in the schools was very much based on the stories of the kings in Heimskringla , and many Norwegians have a vivid impression of the people and scenes of the period from the drawings by famous artists, which were first prepared for the 1899 translation and which have been included in almost all later translations.
Despite its assured reputation as a classic, there has been relatively little research on Heimskringla . Scholarly interest has mainly been concentrated on its textual relationship to earlier sagas or on its trustworthiness as a historical source, in accordance with the tradition of Quel-
lenkunde , inherited from the nineteenth century. Then there have been some literary analyses of style and composition (Nordal, 1973 [orig. 1920]: 132 ff.; Lie, 1937). The question of Snorri's general interpretation of history and his picture of Norwegian society and social change was first placed on the agenda by Halvdan Koht in 1913 (Koht, 1921:76 ff.). Koht regarded Snorri very much like a modern historian with an overall interpretation of the past and explained this interpretation from his experience of thirteenth century society and politics. Koht's approach was followed up among others by Schreiner (1926) and Sandvik (1955). Compared to the discussion caused by Koht's synthesis of Norwegian history in the Middle Ages (chap. 2 below), his interpretation of Snorri seems to have been fairly uncontroversial among historians, though some philologists and historians of literature have voiced their opposition.
Since the 1960s considerable changes have taken place in the study of Old Norse literature. The field has become more international, both in the sense that new milieus have taken part in it and in the sense that there has been increasing interest in analyzing this literature against a European background and studying it with the theories and methods current in other fields. In practice, this had led to widely different approaches: literary analyses according to structuralist or other principles, attempts to reconstruct the social milieu or the mentality of the sagas through theories derived from social anthropology or related subjects, and, not the least, attempts to trace Christian attitudes and European learning in the sagas, in clear opposition to traditional Scandinavian scholarship (cf. Clover, 1985: 239 ff.). However, these approaches have been mainly applied to the Icelandic family sagas and only in a very limited degree to the kings' sagas in general and Heimskringla in particular. The attempts that have been made to interpret Snorri both through theories of primitive mentality (Gurevich, 1969, 1971), Christian philosophy and theology (Weber, 1987, and to some extent Lönnroth, e.g., 1969), or a combination of modern literary and social analysis (Lönnroth, 1973), are mainly sketches. The celebration of the seven-hundredth anniversary of Snorri's birth in 1979 resulted in some new publications (Ciklamini, 1978, and the collection of articles in Snorri: Átta alda minning ) but hardly in new interpretations of Heimskringla as a whole.
On the following pages, I shall try to place Snorri in an international context, in accordance, I hope, with modern scholarship. As a historian, however, my main interest lies in Heimskringla as a description of society and as evidence of social and political attitudes, not in literary or aesthetic aspects. Generally, though recognizing the importance of an author's intellectual milieu and the influence from others, I am inclined to pay more attention to society. This at least seems a natural approach
in Snorri's case, as he was a prominent politician, and Heimskringla immediately gives the impression of dealing with secular politics. Thus, I shall follow in the line of Koht, whose approach I find extremely fruitful and stimulating, though his actual results are evidently in need of revision in the light of recent research over the more than seventy years that have passed since he first presented them.
Heimskringla seems to be little known outside the Nordic countries, despite the fact that it has been translated into the main European languages and that, from a modern point of view—as I intend to demonstrate in the following—it compares favorably with most contemporary works of history both as a narrative and an analysis of history. Thus, it is barely mentioned in the more recent surveys of medieval historiography, such as Smalley (1974), Hay (1977), Guenée (1980), and Schmale (1985). An important exception is Gurevich, who frequently uses it to illustrate medieval mentality and who has also dealt with it separately (Gurevich, 1983; cf. 1969 and 1971).
An examination of Heimskringla according to the principles stated above thus appears like a promising project, the more so as there has been considerable interest in medieval historiography in recent years. As in Europe, in the Nordic countries, the study of historiography was for a long time almost to be identified with Quellenkunde . An important early example of the study of historiography as part of intellectual history, with special emphasis on the basic patterns of thought governing historical narrative, is the work of Bernheim and his school.[1] But the great breakthrough for this approach came after World War II with scholars like Spörl (1965:278 ff., 298 ff.), Lammers (1961), Koch (1965: 321 ff.) and Beumann (e.g., 1965: 135 ff.; 1969). In the Anglo-Saxon world, important contributions were made at a somewhat later date by Brandt (1966), Southern (1970-1973), Green (1972), Smalley (1974), and others.[2] The French Annales school, with its strong emphasis on mentalité in the last two or three decades, has not been particularly concerned with historiography, apart from some important works by Duby, who has used various historiographical works to reconstruct aristocratic ideology (e.g., 1973; 1977: 134 ff., 149 ff.; 1986). The general approach of this school, in emphasizing collective mentality and common patterns of thought and in linking intellectual history to actual social conditions, offers valuable insight. By contrast, the German tradition, often giving excellent analysis of texts, has a tendency to over-emphasize philosophical principles and to isolate the intellectual world from society as a whole.
Notwithstanding my concentration on one particular work by an author who was certainly not representative of medieval man in general, the ultimate aim of my examination is to use historiography to explore
mentality and the connection between culture and society. Medieval historiography contains extensive descriptions of society and politics which are paralleled in no other category of sources and which it must be possible to use, despite the obvious difficulties involved. More specifically, my problem may be outlined as follows.
One aspect of the "civilizing process," which, according to Norbert Elias (1977) and others, Europe underwent from the Middle Ages to the modern period, was the development of "political man" (homo politicus), a parallel to the homo oeconomicus of early capitalism. According to Elias, the warrior of the Middle Ages, who acted spontaneously and according to his impulses and expressed his feelings directly, was substituted by the cold, calculating politician and courtier of the absolutist age, who was used to suppressing his feelings, hiding his opinions, and sacrificing immediate gain for long-term advantage. In a similar way, Max Weber regards "rationalization" as one of the essential features in the evolution of modern society, though in the political sphere he is more concerned with emancipation from religious assumptions than with control of immediate impulses (1964, II: 923 ff., 1034 ff.). To both Elias and Weber these changes have to do with changes in society itself, above all the development of the modern state. Similar general ideas are reflected in much of the histoire de mentalité of the French Annales school, though other aspects of human life than political behavior have been more prominent here. By contrast, the social anthropologist F. G. Bailey (1980) describes the game of politics as being essentially the same in societies as different as the New York mafia, modern Britain, and the Pathans of Pakistan, for example. Despite the strong emphasis on politics in traditional historiography, however, very few historians have been concerned with the general problem of political behavior in the Middle Ages.
For an examination of this problem, the narrative sources are essential. They contain an overwhelming amount of material concerning political behavior. But they do not allow us to observe it directly. They tell us how political behavior was described in the Middle Ages, not necessarily how medieval politicians behaved. There are in principle two ways of dealing with this problem. Traditional political historiography has to a large extent been founded on the implicit assumption that medieval politicians acted more or less as modern ones, and that contemporary descriptions of it, in terms of honor, chivalry, personal friendship or hatred, or religious or moral considerations are the results of deliberate distortions or lack of understanding for what was going on. By contrast, historians of mentalité have been more ready, perhaps too ready, to assume that what is described in the sources corresponds to how medieval politicians actually behaved.[3]
The step from descriptions of political behavior in the medieval sources to actual behavior at the time is thus a difficult one. Moreover, it is impossible to derive very much information on this subject in general from one single narrative source. Nevertheless, it seems important to me to have this problem in mind when analyzing historiographical works, even if the immediate purpose of the present analysis is the more modest one of analyzing descriptions of political behavior.
My book is thus neither an examination of a particular historical problem based on one or more sources, nor a full literary and ideological analysis of one particular text, but something in between. As for Heimskringla as a source, I am not interested in the traditional question of the truth of individual statements, such as the time or place of the battle of
First, there is the problem of the individual author. How can one distinguish between Snorri's personal ideas and those that were common in his milieu? Would it not be more rewarding to take Old Norse literature as a whole or a part of it selected not by author but by subject matter or some other principle and use it as a source for political behavior and ideas of society, without caring about literary form, the author's personal view, or the influence of one text upon another? My answer to this is that Snorri seems to be such an original author and Heimskringla so much of a unity that a fairly detailed analysis of Snorri's authorship is necessary in order to reach definite conclusions on the kind of political game and society presented there. To this, one might object further that if an individual author is to be chosen, a less intelligent one than Snorri would be preferable, because he would be more representative of common attitudes. This is no doubt true, but such authors tend to take so much for granted that it is very difficult for a modern observer to find out what the scattered events they present are really about. An author of Snorri's intelligence and ability to observe, who in addition was an active politician, offers exceptional opportunities of analyzing political behavior and the rules governing it in a medieval society. Admittedly, Snorri may not be representative of contemporary thought and it may often be difficult to distinguish between his personal attitudes and opinions and
those that were common in his milieu. I have no standard method of solving this problem, but I shall try to distinguish between explicit comments in the text and what can be read between the lines, and I will compare him to other, contemporary authors (see chap. 1 below).
Second, there is the problem of using a historiographical work dealing with a period long before the author's own time as a source for contemporary ideas of society and politics: apart from the short section dealing with the very distant past, the narrative of Heimskringla covers the period from the late ninth century till 1177, two years before Snorri was born. Though Snorri probably did not share the modern historicist assumption that the past is qualitatively different from the present (see chap. 5). he may well have imagined people then to have behaved differently from what his contemporaries would do, such as, more heroically or with greater cruelty, or he may have assumed that magic or supernatural intervention was more frequent then.[4] Generally, medieval works of history and not least the sagas, were also works of literature, and this may have determined their presentation of people's behavior. Once more, it is difficult to state a definite method. However, such a long text as Heimskringla offers some opportunities of distinguishing between normal behavior as appears in episodes throughout the work and particularly romantic or heroic stories. A comparison with other saga writers may be helpful in this case as well.
As for combining the examination of Heimskringla as a source for ideas of society and political behavior with analyzing the text in itself, this is perhaps less of a problem than it may seem in the first place. After all, political behavior—in a broad sense—is not one of several themes, it is what Heimskringla is mainly about. An examination of this problem should therefore be able to uncover central aspects of Snorri's historical authorship. This does not mean that it can explain everything in it, nor do I intend to treat every aspect of the work. Philologists in particular may miss one aspect, the analysis of Snorri's sources and his relationship to his predecessors. This is the aspect of Heimskringla and the saga literature that has been treated most extensively—almost exclusively—hitherto, and though far from all problems have been solved, I am not interested in this context in making further contributions to this field and have tried to confine the discussion of such matters to a minimum. As I am equally interested in Snorri's implicit assumptions or "mentality" as in his explicit opinions, I do not always find it important to trace whether he has invented or constructed a particular passage or has taken it from his predecessors. Moreover, it turns out that Snorri is a sufficiently independent mind not to have copied slavishly from his predecessors but usually to have his reasons for omitting or including particular passages. Evidently this does not mean that a comparison between Snorri and his
predecessors is irrelevant. Quite the contrary, it is often of great importance for solving the general problems I have outlined on the preceding pages. I have therefore included comparisons between Snorri and his predecessors throughout the text, whenever I have found it necessary. A summary of these passages and a general conclusion on his way of writing history and analyzing politics as compared to that of his predecessors are to be found in chapter 6.
My intention is thus not to solve the problem of medieval political behavior as outlined above, but to use it as a clue to analyzing Snorri. I hope, however, that such an analysis may make it easier to discuss political behavior in other narrative sources and ultimately in society as it actually was. Though I may be accused of not keeping strictly to the narrow path in choosing my themes, I hope that my readers will find that the material on the following pages is not only there because it suits my interests, but also has some internal unity.
My reason for choosing a work from the outskirts of Europe instead of one of the well-known European historians of the Middle Ages is not to give a representative picture of medieval historiography. Apart from such trivial reasons as Heimskringla being more familiar to me than most European works and a wish to make it more known among modern historians, this choice is made exactly because Heimskringla seems to differ from what is usual in the period and even to differ in the "modern" direction, to some extent anticipating the secular, political historiography of the Renaissance. On the following pages I intend to show this difference and explain the reasons for it.
The book is arranged in the following way. The main problem, political behavior and its relation to society, is treated in chapters 2 to 4, first the conflicts (chap. 2), then Snorri's descriptions of society, human nature, and morality (chaps. 3 to 4). In chapters 1 and 5, on Snorri as an author and a historian, I shall analyze more directly Snorri's own contribution, as opposed to general attitudes in his milieu. Chapter 1 mainly deals with composition, chronology, and Snorri's concept of authorship. The reason for dividing this from chapter 5 and placing it in the beginning of the book is that it is also intended as a general presentation of the contents of Heimskringla , placing the various stories, which are treated more in detail in the following, in their appropriate context. In chapter 5, I shall then try to show how Snorri's ideas of the game of politics determine his general interpretation of history, which differs on important points from the main trend in contemporary European historiography. Finally, in chapter 6, I shall compare Snorri with the historiographical tradition in Norway and Iceland and in Europe in general and try to explain his particular approach in terms of social and cultural conditions in Norway and Iceland. The rest of this introduction contains a general
presentation of the work and its author and sketches of Norwegian medieval history and the evolution of Old Norse historiography. This is primarily intended for readers who are not familiar with Old Norse studies.
The General Background: Snorri and Norwegian History
Heimskringla deals with the history of the Norwegian kings, mainly from the late ninth till the late twelfth century. It thus covers the period of the formation and early development of the Norwegian monarchy. The "story" of this, as told by modern scholars, runs approximately as follows.[5] Norway was united under one king for the first time by Haraldr hárfagri (c. 890/900-945), who conquered most of the country. The decisive battle, celebrated by scalds and saga writers, took place in Hafrsfjord, near what is now Stavanger, probably shortly before 900. After Haraldr's death, various pretenders sought to take over his position but no one succeeded completely. The country remained divided in several lordships, though fewer and larger than before Haraldr. There were frequent internal struggles, in which the Swedish and particularly the Danish king often intervened. For long periods, the Danish king even controlled parts of the country. The external and internal struggles of this period were intimately connected with the Viking expeditions, most of the pretenders having enriched themselves in this way or having served under foreign kings.
One of these warlords, King Óláfr Haraldsson (1015-1030), later known as St. Óláfr, managed to unite the country once more and made Christianity the official religion. Óláfr was later defeated and killed in the battle of Stiklestad (1030) by his internal opponents in alliance with King Cnut the Great, who ruled both Denmark and England. Nevertheless, in a long-term perspective his reign made a turning point. During the following century, Norway remained a united monarchy, though for some periods ruled by several kings as co-rulers. Its independence was maintained, and sometimes its kings even conducted aggressive wars against Denmark or the British Isles. Christianity became firmly established and the ecclesiastical organization was built up, until an independent Church-province, led by the archbishop of Nidaros, was established in 1152/1153. This organization also served to give the king a firmer hold on the country. Further, the lay magnates now became more closely attached to the king, thus furthering the process of centralization.
This development was apparently broken off in the 1130s, when a disputed succession inaugurated a series of civil wars which went on intermittently until 1240. Actually, this period was an important phase in the
centralization of the country, and the wars themselves to some extent contributed to this. They seem to have started as struggles between individual rulers and their friends and clients and then gradually developed into prolonged wars between permanent parties, which controlled different parts of the country. In the 1160s Erlingr Ormsson, nicknamed skakki (crooked-neck), assembled the leading churchmen and the majority of the aristocracy around the kingdom of his son Magnús. His opponents increasingly had to seek support from marginal groups or at least the less prominent members of society. One of these opponents, however, Sverrir Sigurðsson (1177-1202), managed, largely through his own political and strategic ability, to fight his way to the throne and defeat the party of Erlingr, though he had to fight various rebellions until his death. Under Sverrir and his successors, a new aristocracy emerged, to a considerable extent recruited from below and more closely attached to the king. Sverrir and his adherents had considerable problems with the Church, to some extent also with the old aristocracy. They developed a "party ideology," which was anticlerical, "democratic"—in the sense that it encouraged recruitment from below to political positions—and strongly monarchical. This remained the official ideology of the dynasty for a long time, though in a somewhat diluted form. The dynasty of Sverrir remained on the throne, though they had to strike a compromise with the leaders of the opposite party. In the years following Sverrir's death the old and the new aristocracy, to some extent also the Church, increasingly became united under a strong monarchy. Norway then entered its "age of greatness" (1240-1319). This was a period of strong monarchy, administrative centralization, internal peace, cultural achievement, and a relatively strong external position.
Snorri's Heimskringla was written toward the end of the period of civil wars, when the dynasty of Sverrir was fairly firmly established on the throne. According to some scholars, Snorri's general interpretation of history was influenced by contemporary or recent Norwegian conditions, notably the conflict between monarchy and aristocracy. He thus tells largely the same story as that of modern historiography, though in the guise of dramatic events. As I shall develop later, I do not share this opinion.
Snorri begins his history with the pagan God Óðinn, the mythical founder of the dynasty. In contrast to Saxo Grammaticus in Denmark, who devotes a major part of his work to the "prehistory," Snorri dismisses the early period rather briefly. The kings from Óðinn, according to Snorri a contemporary of the Roman conquerors (Yngl . chap. 5), until the mid-ninth century, are grouped together in the Ynglinga saga , which is little more than an extended genealogy. The reason for this may have been lack of information, to some extent also doubts of the reliabil-
ity of the sources that were available. The real narrative of the Norwegian kings start with Hálvdan svarti, the father of the great conqueror Haraldr hárfagri. Heimskringla ends early in 1177, when King Magnús Erlingsson won what seemed the final victory over the last rebel faction, the Birchlegs, in the battle of Ré. This was immediately before the rise of Sverrir, who became the leader of the remains of the defeated army shortly afterward. Snorri's reason for ending his story here was probably that the following events had already been treated in detail in Sverris saga . From a modern point of view it seems curious that he does not in any way anticipate Sverrir's rise to power. If we did not know better, we would have thought that the battle of Ré marked the final victory for the dynasty contemporary to Snorri.
During most of the nineteenth century, Snorri was considered the most important source for the early history of Norway. In 1911 the Swedish historian Lauritz Weibull published a series of "critical examinations" on early Nordic history in which he rejected most of the information given by Snorri and other Icelandic historians on Nordic history of the tenth and eleventh centuries, demanding that all conclusions should be based on contemporary evidence or on later evidence only if it could be controlled.[6] This meant that most of what one believed to know about this period was thrown away in a single stroke. Weibull's book created violent disputes, but his ideas gradually established themselves as the orthodox opinion, at least in Denmark and Sweden, to some extent also in Norway. In 1931 the Norwegian historian Edvard Bull declared:
We . . . have to give up all illusions that Snorri's mighty epic bears any deeper resemblance to what actually happened in the period between the battle of Hafrsfjord and the battle of Ré. (Bull, 1931: 9)
Bull was more skeptical than most, but his statement is nevertheless an indication of a general trend. Historians are reluctant to use Snorri as a source, at least for events before the twelfth century, though they have not been very explicit concerning what to trust and what not. The problem is that there is little else to use. There is some archeological evidence. As for written materials, the earlier sagas that are extant are not much older than Heimskringla and were mostly also Snorri's sources. From around 1150 there is some documentary evidence, above all papal letters and other ecclesiastical sources. Foreign sources, such as the Anglo-Saxon Chronicle (Earle and Plummer, 1892-1899) and Adam of Bremen's Gesta Hammaburgensis ecclesiae pontificum (c. 1070) contain some information on the tenth and eleventh centuries. Apart from this, the main source of information is the scaldic poems, mostly laudatory addresses to kings, which are quoted extensively by Snorri and other
historians. Because of their complicated metrical form these poems are believed to have been preserved orally over the centuries, though the possibility evidently exists that some of them may be later inventions. They contain some basic information on the names of kings, battles, and so forth, but no real narrative. Finally, we know that Snorri used written sources from the early twelfth century which are now lost and from which he probably derived some basic information concerning main events, deaths of kings, and so forth.[7]
Bull's statement concerning Snorri has to do with trustworthiness but also with relevance. Like most early historians, Snorri describes numerous battles and dramatic events, in which modern historians take little interest, while omitting what they find most interesting, such as government, administration, social and economic conditions, and the material necessary to trace the main lines in the development of society. Even when dealing with political events, he is often vague concerning the exact contents of treaties, the claims of the parties in a conflict, and so forth. In short, both because of increased skepticism concerning the trustworthiness of Snorri and other saga writers and because of the development of modern Norwegian historiography away from history of events to history of social structure, Heimskringla has largely lost its dominant position in Norwegian historiography. At the same time, this historiographical development has opened new possibilities for using Snorri as evidence for thirteenth-century attitudes to man, society, and history.
Snorri's Life and Authorship
Heimskringla is named after its opening words Kringla heimsins (the circle of the world). The text has been preserved in several medieval manuscripts and fragments, most of which are Icelandic. What is considered the most important manuscript, Kringla , written c. 1250-1280, was destroyed in the fire in Copenhagen in 1728, except for one leaf, which had been brought to Stockholm and is still there. Its text, however, is known from seventeenth-century transcripts (Lie, 1961: 302; Louis-Jensen, 1976: 16 ff.). No extant manuscript of Heimskringla names Snorri as its author before the translations by the Norwegian antiquarians Laurents Hanssøn (1550) and Peder Claussøn Friis (1599). However, there seems to be good reasons to believe, as most scholars do, that this attribution stems from medieval manuscripts that are now lost.[8] This hypothesis is supported by casual references in other sagas to information derived from Snorri, which is actually to be found in Heimskringla . It thus appears that we may fairly safely consider Snorri the author of Heimskringla . In contrast to most other works of the Old Norse litera-
ture, with Heimskringla we then find ourselves fortunate in not only being able to name the author, but also being faced with an author about whom very much is known. Snorri was one of the mightiest chieftains of thirteenth-century Iceland, and the contemporary Íslendinga saga , written by his nephew Sturla Þórðarson (in the collection Sturlunga saga ), contains much information on his life and political activities.
Snorri was born at Hvammur in Western Iceland in 1179.[9] He had prominent ancestors on both sides, but it was his father, Sturla Þórðarson the Elder, who brought the family into the foremost rank of the Icelandic aristocracy. Snorri, however, was not immediately able to take over his father's position. He was the third son and only four years old at his father's death in 1183. By the time he had grown up, his mother had wasted most of his fortune. But Snorri had a great advantage in being fostered by Jón Loptsson of the Oddaverja family. This family, at the time the mightiest of the country, ruled a large part of Southern Iceland. The time Snorri spend at Oddi, until Jón's death in 1197, was also important from another point of view. Oddi was a center of learning and historical studies,[10] and Jón was a learned man; he was related to the Norwegian royal family—his mother was the daughter of King Magnús berfoetr—and had personal memories of Norwegian history from the years he had spent there. Snorri may thus have received general education and inspiration for his future historical studies at Oddi, besides direct information on past events from Jón.
Thanks to his connection with the Oddaverjar, Snorri was able to contract a very rich marriage in 1199, which laid the foundation for his future prominence. In the following years he increased his power by collecting goðorð (chieftainties) and staðir (churches),[11] particularly in the region around Borgarfjörður, thus furthering the process of power concentration which during the first half of the thirteenth century divided the country into a few large territories (ríki), governed by mighty chieftains. In the years 1215-1218 and once more 1222-1231 he was
practice, however, his Norwegian connections mainly served to further his own interests as an Icelandic chieftain. The last period of his life, after his first return from Norway, coincided with one of the most dramatic periods of internal strife in the history of the Icelandic free state, and Snorri played a prominent part in these struggles. In 1237 Snorri temporarily withdrew from the internal strife in Iceland to spend two years with his friend Skúli in Norway, but this involved him in another struggle, the one between Skúli and King Hákon, which led to Skúli's rebellion late in 1239. King Hákon now considered Snorri his enemy, and their relations deteriorated further when Snorri left Norway against the king's express orders in 1239. The king now ordered his chief representative in Iceland, Gizurr Þorvaldsson, to take Snorri captive and return him to Norway, or kill him. Gizurr had been married to Snorri's daughter, but the marriage had been dissolved. Gizurr allied himself with another former son-in-law of Snorri, Kolbeinn ungi, opened a lawsuit against Snorri, and used arms to further his claims. In September 1241 he attacked Snorri at his farm Reykjaholt and had him killed.
Snorri's political career thus ended in failure and death. As a politician he has generally not been regarded highly by modern historians. According to common opinion, he was intelligent, ambitious, probably not very courageous in war, and behaved with a curious mixture of hardness and weakness, which seems to cover a fundamental uncertainty. No one, however, has disputed his ability to analyze politics. There may be some connection here: the politician's determination and ability to reach sudden decisions and the historian's reflection are not easily combined. However, much remains to be done before we have a complete picture of the internal strife in Iceland during these years and the basic conditions of politics in the last phase of the Icelandic free state, and this makes it difficult to form a definite opinion on Snorri's ability as a politician. For the present purpose, I shall confine myself to pointing out that Snorri both had considerable political experience and good connections to Norway and the Norwegian court. The relevance of this for his work remains to be examined.
Apart from Heimskringla , Snorri is also, according to a manuscript dated around 1300, the author of the Younger Edda , a handbook on scaldic poetry, which is an important source for Old Norse mythology and for contemporary theory of language (see most recently Clunies Ross, 1987; von See, 1988). He is also the author of the Separate Saga of St. Óláfr . The former is probably a comparatively early work, for the most part written before Snorri's first journey to Norway, though it cannot have been finished until after 1223. The Separate Saga is largely identical with the Óláfr's saga of Heimskringla and is usually considered to have been written earlier. Further, most scholars attribute one of the
most famous of the Icelandic family sagas to Snorri, Egils saga Skallagrímssonar . This has usually been considered an early work, though Jónas Kristjánsson has suggested a date after Heimskringla (1977: 470 ff.). There appears to be universal agreement that both the Separate Saga of St. Óláfr and Heimskringla were written after Snorri's first journey to Norway. Though Snorri may have received much of his historical information in Iceland, the geographical accuracy of both works strongly indicates an author who had visited large parts of the country. This also applies to Egils saga . The conventional dates are c. 1220-1225 for the Separate Saga and c. 1230 for the Heimskringla . The latter date is suggested by a casual reference in Sturlunga saga to a visit to Reykjaholt by Snorri's nephew Sturla Sigvatsson, during which he made transcripts of saga books that his uncle had composed. The period c. 1230 was also a peaceful interlude in Snorri's struggles with his rivals, which may have offered an opportunity for literary activities. The exact date of these works, however, is of slight importance for our evaluation of Snorri as a historian. We know fairly well that both were composed when he was a mature man, with long experience as an Icelandic politician of the first rank and with good knowledge of Norwegian conditions and connection with the Norwegian court. As his works deal with the fairly distant past, it is doubtful whether we can expect them to be influenced by particular events in his own life, though we cannot quite rule out the possibility that varying attitudes to the Norwegian monarchy may be reflected in them.
Snorri's Background and Sources
Heimskringla represents the climax of Old Norse historiography in a double sense: it comes at the end of about half a century of extensive saga writing and includes material contained in most earlier works dealing with the Norwegian kings. It is also generally considered the best and most mature example of a kings' saga, both from a literary and a historiographical point of view. Its sources have been dealt with extensively by a number of scholars. There has also been considerable discussion on the origin of the saga literature.
The two historiographical traditions of contemporary Europe, the clerical and the aristocratic (see Brandt, 1966), have their parallels in Old Norse historiography. It is customary among students of language and literature to distinguish between the "learned" and the "popular" style (Halvorsen, 1962: 119 ff.), the former being influenced by Latin and usually the medium of clerics, whereas the latter corresponds to local traditions and is the medium of the laity and the classical saga literature. In historiography the learned works usually but not always
reflect religious and clerical points of view and generally bear close resemblance to European clerical historiography. The other tradition shows some similarity to European aristocratic historiography, but is more original. As its specifically aristocratic features are less prominent the term "secular" is more suitable than "aristocratic."
There is a solid tradition in scholarship for this distinction and for stressing the unique achievement of Icelandic saga writing in contrast to the learned and clerical culture of contemporary Europe, particularly within the so-called Germanistic school. In the last decades, however, there has been a tendency in the opposite direction, stressing the basic unity of medieval culture and attempting to interpret the sagas in terms of European, Christian philosophy and theology (e.g., Lönnroth, 1964, 1965, 1969; Pálsson, 1971, 1984, and 1986; Harris, 1986; Weber, 1987, see also Clover, 1985: 264 ff.). To some extent, this has been a healthy reaction against much obscure Germanic romanticism of the last century and has also been stimulating in attempting to link the North to the common European culture. Generally, however, I think this reaction has gone too far, not the least because it tends to minimize the difference that existed between the two cultures in Europe as a whole (cf. also von See, 1988: 13 ff. and passim ). I shall touch upon this problem repeatedly in the following pages and confine myself to a few remarks regarding Snorri's predecessors in this context.
The first historical works dealing with the Norwegian kings by the Icelanders Sæmundr and Ari, which date from the early twelfth century and are now lost, cannot be easily fitted in with either category. They were apparently short, dry summaries of events, their authors exerting themselves in establishing an exact absolute and relative chronology. This is the impression we get from Ari's Íslendingabók , which is now extant. His Konunga ævi (kings' lives) was most probably an appendix to Íslendingabók , which was removed later, though some scholars have maintained that it was a separate work (see Ellehøj, 1964: 43 ff. with ref.). These works lacked both the vivid narrative of the secular saga and the religious interpretation of the clerical one. The fact that both these authors were priests is of less relevance in this context, as Sæmundr certainly and Ari probably combined his priesthood with the position as a chieftain. This combination was quite usual in eleventh- and early twelfth-century Iceland. Sæmundr apparently wrote in Latin, whereas Ari wrote in the vernacular. Both must have influenced a number of later historians, but the extent of this influence is difficult to detect and hotly disputed.[12]
The so-called Norwegian synoptics represent the next stage in the development, as far as extant works are concerned. The three works that can be classified under this heading, Theodoricus Monachus's Histo-
ria de antiquitate regum norwagiensium (c. 1180, in Latin), the anonymous Historia Norvegiae (before 1178, c. 1220, also in Latin), and Ágrip af Nóregs konunga sogum[*] (c. 1190, anonymous, in Old Norse) are written from a distinctly clerical point of view. The milieu of Historia Norvegiae is unknown, whereas the two others show connection with the archiepiscopal see of Nidaros. The oldest history of St. Óláfr, Passio Olavi , written by Archbishop Eysteinn Erlendsson (1161-1188), also belongs to this milieu, but is better classified as a saint's legend than as historiography in the real sense. The clerical point of view of these three works is expressed both in their ideology, classifying the kings according to the Augustinian schema of the rex iustus and rex iniquus , their frequent appeals to God's intervention, and their concern with the moral aspect of human acts. As in many contemporary European clerical chronicles and in contrast to the classical saga, their narrative is mostly dry. Many of the dramatic episodes that occur in the later sagas are omitted, and those included are told in a short and undramatic way.[13] Ágrip bears some resemblance to the classical sagas in that its author remains in the background and that it contains some vivid stories, told very much in the classical saga style. In the two Latin works, however, the author comes forward explicitly, commenting on the events. The chief interest of these authors is clearly in interpretation, not in narrative. They want to uncover the moral and religious significance of the events (Bagge, 1989a ). Characteristically, they reserve rhetorical embroidery to expository and interpretative passages.
The birth of secular historiography, the classical saga, must be roughly contemporary with "the Norwegian synoptics." The secular saga was once believed to have developed gradually from clerical historiography. The "intermediate forms," represented by works like the Oldest Saga of St. Óláfr, which is preserved only in fragments, and the Saga Óláfs Tryggvasonar by Oddr Snorrason munkr, were considered evidence of this. Recent research having both questioned the connection between the Oldest Saga and hagiography and the early date of the work, no direct evidence remains that the "intermediate forms" represent an early stage in the development.[14] Nor does this seem very likely on a priori grounds. The two forms are so different that it hardly seems a likely explanation that the latter should have developed from the former over a period of about twenty years.
It is very difficult to detect early forms of secular saga writing. Most works are difficult to date, and those that can be shown to be early with reasonable certainty, like the second and larger part of Sverris saga and Orkneyinga saga , both probably written shortly after 1200, represent saga art in its full development: the characteristic short, terse "saga style," the vivid narrative, the pre-dominance of relative as opposed to
absolute chronology, the secular attitude, and the use of speeches and dialogue for purposes of analysis and interpretation, while the author remains in the background. Concerning this latter point, however, some development may be detected, both in the various versions of Oddr munkr's Saga Óláfs Tryggvasonar and in the difference between the first part of Sverris saga (Grýla ), which was probably written already in the 1180s, and the rest (Holm-Olsen, 1977: 58 ff.; 1987: 79 ff.). In the discussion of sources of the extant sagas, a number of earlier sagas, mostly separate sagas of individual kings, are thought to have existed before the extant ones. If this is the case, the origin of the secular saga is to be found some time back in the twelfth century. This is not very helpful, however, since nothing is known of the contents and style of most of these works. As a matter of fact, with a few exceptions, their very existence remains doubtful.
The most important exception to this is Hryggjarstykki , the existence of which is attested both by Morkinskinna and Heimskringla . This work was written by the Icelander Eiríkr Oddsson, probably in the 1160s, and covered either the whole or parts of the period 1130-1161. If the former hypothesis is correct and the detailed and very similar treatment of the wars between Haraldr gilli's descendants 1155-1161 in Morkinskinna, Fagrskinna , and Heimskringla stem from Hryggjarstykki , then we have evidence of the fully developed secular saga as early as the mid-twelfth century. The most recent scholar who has treated the matter, Bjarni Guðnason, denies this, however, and maintains that the work was a biography of the pretender Sigurðr slembir and only covered the years of his struggle for the throne, 1136-1139 (Guðnason, 1978). If this is correct, there may be some evidence for its dependence on a hagiographic tradition and thus for the idea of the secular saga developing from clerical historiography. Sigurðr was tortured to death by his enemies and bore his pains with the utmost tranquillity of mind, chanting psalms all the time. The description of this clearly resembles a story of a martyr. To what extent the story of Sigurðr can also be understood as a description of a secular hero, remains to be considered (see later discussion). It may be mentioned in the same context that Ludvig Holm-Olsen suggests a similar model for another saga, the earlier part of Sverris saga (1953:91 ff.). However, Guðnason also points to other influences behind the work, notably oral storytelling, as is described in the Sturlunga saga in connection with the wedding at Reykjahólar in Iceland in 1119 (Sturl . I: 18 ff.; see, e.g., Jónsson, 1923: 198 ff.).
Even if we accept the idea of influence from hagiography on the description of Sigurðr's death, it is difficult to see that this explains very much of the distinctive features of the classical saga. First, we cannot exclude the possibility that descriptions of men like Sigurðr and Sverrir
may equally well have been derived from stories of secular heroes. Second—and most important—most of the distinctive traits of the classical saga are conspicuously absent from saints' lives, namely the characteristic saga style, the retreat of the author, the speeches and dialogues used as interpretation and comment, and the description of the heroes. For instance, the vitae of the saints almost never describe their heroes' appearances, whereas the sagas usually do. The descriptions of character are also very different, both in style and content. The "hagiographic explanation" thus seems to be no explanation at all. It may possibly explain some features of individual works and even to some extent—though this is more doubtful—why they were written, but it is unable to account for the characteristic features of the classical saga. And this is after all the whole problem.
Thus, the question of the origin of the saga still lies in the dark. The problem is difficult, perhaps impossible to solve. But it must be pointed out that the sort of research that has been practiced hitherto is not very likely to lead to its solution. In accordance with the bookprose theory, which for a long time has been the prevailing orthodoxy in saga studies (see Clover, 1985: 241 f.), scholars have mainly been interested in tracing the origin of extant texts back to other, earlier texts, extant or not. This approach must sooner or later lead to a dead end. How can the earliest text be explained? In general, it is hardly possible to explain new developments in literature simply by comparing texts. Social and intellectual milieus must be taken into account, and this has not been done very much in scholarly research on saga origins.
Though it is outside the scope of the present work to discuss the question of saga origins and the milieus behind the first sagas in general, something must be said on the relative importance of the Norwegian royal court and the aristocratic milieus of Iceland. Most of the sagas of this early period seem to have been written by Icelanders. In contrast to the Icelandic family sagas, which are mostly later, they deal with matters outside Iceland, but by no means exclusively confined to Norway. There are sagas of the Orkney earls, of the chieftains of the Faroe Islands, the Danish kings, and the jómsvíkingar (the Vikings of Jom[*] , i.e., the island of Wollin in the Baltic), and there are indications that there were also sagas of the earls of Lade and their kinsman Hákon Ívarsson. Thus, the kings' sagas were not necessarily the result of Norwegian royal patronage. There seems to have been a general interest in such stories within the Icelandic aristocracy, which in some cases, though by no means always, may have to do with kinship ties with the persons or families described in the sagas.
The sagas of Norwegian kings that belong to this period are Oddr Snorrason's and Gunnlaugr Leifsson's sagas of Óláfr Tryggvason, the
Oldest Saga and the Legendary Saga of St. Óláfr, and the earlier part of Sverris saga , the so-called Grýla . This latter work was written under the king's supervision by the Icelandic abbot Karl Jónsson, and is thus clearly an example of royal patronage.[15] Whether the original initiative was Karl's own or Sverrir's is a debated question, as is the question of the initiative and milieu behind the latter—and larger—part of the saga, which was written after Sverrir's death in 1202.[16] The two sagas of Óláfr Tryggvason are both composed by Icelanders but originally written in Latin. Oddr's is preserved in translation, whereas Gunnlaugr's is lost but can partly be reconstructed from Óláfs saga Tryggvasonar hin mesta , a compilation from around 1300. These works seem to represent an intermediate form between clerical and secular historiography, combining a distinctly clerical outlook with vivid narrative. Their particular interest in King Óláfr Tryggvason has been explained as an expression of the wish to emphasize the king who brought Christianity to Iceland against his more important namesake St. Óláfr (Lönnroth, 1963: 54 ff.). There may thus have been a specifically Icelandic reason for writing about this particular Norwegian king. By contrast, Norwegian interests are likely to have been more prominent in the case of the two sagas of St. Óláfr. St. Óláfr was prominent in the propaganda of both parties in the ideological warfare between King Sverrir on the one side and King Magnús Erlingsson and his descendants on the other, the latter for a long time receiving support from the Church. Though the works in question can hardly be regarded as direct propaganda for either party in this conflict, it is not unreasonable to see a connection between Óláfr's political importance at the time and historical interest in his life.[17] Whether this also applies to the slighty later saga of St. Óláfr by the Icelander Styrmir Kárason, which is only preserved in fragments,[18] or whether this is a product of an Icelandic milieu, is more difficult to tell.
The following period, approximately 1220-1235 is the period of the great compendia on the Norwegian kings and in general a period in which Norwegian matters dominate saga writing. These works are, in addition to Snorri's Separate Saga of St. Óláfr and Heimskringla , the anonymous Morkinskinna (c. 1220) and Fagrskinna (c. 1225). Of these sagas, Fagrskinna is clearly written in close connection with the Norwegian court, possibly also by a Norwegian,[19] whereas the rest are commonly considered to have been written by Icelanders. Their connection with Norway and the Norwegian court remains doubtful. Are they a continuation of the Icelandic interest in heroes and chieftains of other countries in the North of the preceding period, or is some kind of Norwegian patronage behind them? Arguments in favor of the latter point of view are the undoubted interest in history and intellectual work in general at the contemporary Norwegian court as well as the fact that the
political links between the Norwegian king and the Icelandic chieftains became closer as the result of the attempts by the Norwegian king to bring Iceland under his dominion. As we have discussed, Snorri had close connections with the leading circles of Norway and became involved in these attempts. Norwegian influence behind the conception and execution of the work cannot therefore be excluded, but the matter must ultimately be decided by internal evidence. The question is more difficult to decide in the case of Morkinskinna , as we know nothing of its author, and there has been little research on the work. But we may note that it contains many more stories of Icelanders than the two other works.
Morkinskinna deals with the period from the death of St. (Óláfr till the death of King Eysteinn Haraldsson, that is, 1030-1157, but may possibly have extended till 1177 (the battle of Ré). It is more loosely structured than the two other works, often consisting of rather disconnected stories, and there are clear indications that it has been reworked and interpolated. Fagrskinna covers the period from Hálfdan svarti in the middle of the ninth century till the battle of Ré (1177). It is the shortest of the three and also the least saga-like. Despite its secular attitude, it resembles the clerical works in its short and often dry narrative and in omitting many dramatic stories, above all those dealing with internal strife, to some extent also in its rex iustus -ideology (see later discussion). Together with Heimskringla these two works represent the kings' saga writing in its full maturity. On most points, however, it is difficult to trace a development compared to the earlier works, though it can be done concerning scaldic poetry. The more frequent and conscious use of this clearly has to do with the fact that these works dealt with a remote period, from which few sources were available.[20] Generally, Heimskringla clearly represents progress compared to its two immediate predecessors. This has been frequently pointed out by scholars and will be demonstrated in the following.
The relationship between these sagas mutually and with the earlier saga tradition has been the subject of considerable learned discussion, which need not detain us here. From our point of view the sources of Heimskringla are the most important aspect of the question. Snorri mentions some of his sources explicitly, such as Ari and Eiríkr Oddsson, but most of them must be reconstructed by comparison with other sources. There is general agreement that he used Morkinskinna , and most scholars also think that he used Fagrskinna for Heimskringla , though not for the Separate Saga . As for earlier sagas he has evidently used Oddr Snorrason's Óláfs saga Tryggvasonar . The question of the sources for his biography of St. Óláfr is more complex, but he used most
of the stories contained in the Legendary Saga , however it is unknown whether he knew this work directly or found the same material in a source which is now lost, such as Styrmir's history. Coming at the end of an extensive, if not particularly long tradition, Snorri thus seems to have borrowed from most of his predecessors.