ABSTRACT: Scholars and readers have long been interested in the historical validity
of the saga literature. This study addresses how the notion that the sagas had historical
value was rationalized throughout the 20th century and goes on to explore how some
late 20th-century questions about historical validity, well-known in the humanities
at large but rarely asked in saga studies, might cast some light on the possibilities
and impossibilities of finding historical truths in the sagas.
RÉSUMÉ: Les chercheurs et les lecteurs se sont depuis longtemps intéressés à la validité
historique du genre littéraire des sagas. Cette étude traite de la façon dont la notion
selon laquelle les sagas ont une valeur historique a été rationalisée tout au long
du XXe siècle, puis se penche sur la façon dont certaines questions de la fin du XXe
siècle sur la validité historique, bien connu dans les sciences humaines en général,
mais rarement abordées dans les études des sagas, pourraient mettre en lumière les
probabilités et improbabilités de découvrir des vérités historiques dans les sagas.
In a recent survey of early medieval England and Arthurian legend, Guy Halsall gave
short shrift to those who wish to believe in a “real” King Arthur, stating that “our written evidence is absolutely incapable of proving that Arthur existed, and certainly
of telling us anything reliable about him,” although he was careful to add that “its faults do not prove that he did not exist” (Halsall 86). The legends of King Arthur can be studied as such, but do not yield any proof that
this figure existed as he is depicted in the legends. Contemporary sources provide
three historical Arthurs but offer no indications that these are the basis of the
Arthur of legend. The warlord Arthur first appears as a somewhat nebulous figure in
9th- and 10th-century sources and the explosion of Arthurian legend only begins with
Geoffrey of Monmouth’s Historia Regum Britanniae in the 12th century. Thus, although King Arthur has been a celebrated cultural hero
for centuries, there is no “historical” King Arthur behind the legends and he cannot be resurrected from textual sources.
Halsall’s study offers some revealing examples of a kind of modern pseudo-history
fueled by the desperation to find the truth behind the legend (Halsall 137–54). The horror vacui generated by “dark ages” of history where no reliable contemporary sources can provide us with comfortable
truths is clearly not significantly in retreat, and is often met with scholarly invention
by both amateurs and scholars alike. Indeed, the King Arthur situation, wherein zealous
scholars try to satisfy this need for “the truth behind the legend” by distilling “facts” from dubious and much younger sources, is far from unique. In this study, I will
examine a similar aspect of certain 20th-century scholarly attitudes towards the Old
Icelandic sagas of the Late Middle Ages as a much-needed source of “factual information” in many a general history, handbook, and encyclopedia.
In the past these sagas also served as sources of information about the Germanic
and Scandinavian past (500–800) as well, although the information that Old Norse textual
sources from the late Middle Ages provide about Attila the Hun and Rollo was sometimes
regarded as fabulous or legendary as early as the late 19th century.
However, as sources for the 9th and 10th centuries, the sagas have continued to be
regarded as having historical value. The primary aim of the present study is to examine
how this notion was rationalized throughout the 20th century and to explore if some
late 20th-century questions about historical validity, well-known in the humanities
at large but rarely asked in saga studies, might cast some light on the possibilities
and impossibilities of finding historical truths in the sagas.
Discovered by the humanist scholars of the 17th century and attracting great interest
from the learned, leading to the great “manuscript war” between Denmark and Sweden and the first publications of sagas as “sources” for ancient Scandinavian history, kings’ sagas and sagas of Icelanders were treated
as factual sources well into the 20th century, although skepticism increased during
the empirical 19th century.
It may be said to have culminated in the critical survey of the use of sagas as sources
for the political history of Scandinavia in the late 10th and early 11th century by
the Swedish historians and brothers Lauritz Weibull (1911) and Curt Weibull (1915), so critical in their respective approaches that each man sternly included the word
“critical” in his title. The Weibull brothers’ criticism clearly called for a dramatic new re-evaluation
of the sagas as historical sources.
Some scholars, including Finnur Jónsson (1912), continued to insist that the sagas could still be used as factual sources, pointing
to the skaldic poetry that appears within the sagas as contemporary and thus more
likely to provide accurate factual information than the sagas themselves.
To this day some scholars and others involved with the presentation of history, particularly
in Iceland, will insist that, though not the best possible sources, the sagas can
still be mined for factual information about the 9th and 10th centuries. In Iceland,
for example, the legendary 9th-century settler Ingólfr Arnarson is often still spoken
of as a “real” and “historical” figure,
even though the formal characteristics of the 13th-century version of his legend
that have much in common with other foundation myths were long ago identified by Sørensen.
Other historians clearly grew uneasy about the purported source value of the sagas
and felt it necessary to rise to the challenge and adapt their approach to befit a
more critical era. One of the first into the breach was the Norwegian historian Halvdan
Koht (1914) who urged a new focus on the kings’ sagas as authorial works from the 12th and 13th
centuries and emphasized their relationship with their time of composition. Koht suggested
that the sagas should be categorized based not only on their alleged source value,
the foundation for the popular classification of various sagas into king’s sagas,
contemporary sagas, sagas of Icelanders, legendary sagas, etc., but also based on
their historical philosophy and their “partistilling” [political stance]. For Koht, this meant a classification of the kings’ sagas into groups based on their
respective reflections of either aristocratic, royal, or clerical points of view.
These categories may now seem slightly banal, and have indeed been followed by several
saga studies in which scholars refashioned the sagas as mere propaganda pieces, wholeheartedly
hostile towards certain institutions or individuals whilst favoring others. Still,
Koht’s study marks a dramatic shift in the concerns of saga scholars: hitherto concerned
mostly with the “saga age,” the age of the events depicted in the sagas, scholars now increasingly turned their
focus towards the age during which the sagas were composed.
In the decades following Koht’s study, a new approach in saga studies was developed,
one in which the focus was increasingly on the saga authors rather than the heroic
figures in the sagas, culminating in the event that heroic warriors were now replaced
by heroic artists (Helgason 1998). This diminishing interest in the historical value of the sagas may have peaked with
Sigurður Nordal’s study of Hrafnkels saga (Nordal 1940) in which he shifted that particular saga from one absolute category, history (reality;
truth), to what he regarded as its antithesis, fiction (art; literature). Hrafnkels saga was, in Sigurður Nordal’s view, unhistorical and that made it fictitious. However,
he dissented strongly from the hierarchy established by 19th-century scholars wherein
fiction was inferior to history and argued that Hrafnkels saga, while not very historical (meaning accurate), was an important work of art, and
perhaps all the more valuable for it.
Thus in the mid-20th century, the sagas had escaped the ignoble fate of becoming bad
history and were instead transformed into fine art. And yet Sigurður Nordal and his
contemporaries never completely abandoned the old ideas pertaining to the historical
value of the sagas, exemplified in the Íslenzk fornrit editions published between 1933 and 1959.
In 1933, Egils saga appeared as volume 2, and yet the first volume to see the light of day, in the now
standard Íslenzk fornrit series. While Einar Ólafur Sveinsson seems to have undertaken a considerable amount
of work on this edition, Sigurður Nordal is the sole credited editor of the volume
and also contributed an introduction where he, amongst his other concerns, discusses
the saga’s relationship with history, its relationship with other “sources,” and, in great depth, the historical chronology of the sagas (Sigurður Nordal 1933, xxxvi–liii). The latter discussion is based mostly on the description of the battle of Vínheiði
that appears in the saga and the actual Battle of Brunanburh in 937, mentioned in
the Anglo-Saxon Chronicle.
Sigurður Nordal’s introduction includes a chronology of the life and times of Egill
Skalla-Grímsson, the eponymous hero of the saga, from his birth around 910 to his
death around 990. The inner chronology of the saga does not quite work (which is the
case for most of the sagas) so it is adjusted to accord properly with known facts
from older and more reliable sources (which nevertheless do not mention Egill) such
as the Anglo-Saxon Chronicle. Thus every reader of this edition of Egils saga is informed that Egill went to York to meet King Eiríkr and composed the poem Hǫfuðlausn
in the year 948: “Þó að sögunni kunni að skjátlast þar um einstök atriði, er sjálfur atburðurinn studdur
órengjandi heimildum” [Though the saga may be wrong in some of the details, the event itself is verified
by irrefutable sources] (Nordal 1933, xlvii). The reader is thus left with little doubt that Egill was an actual historical person.
That does not merely mean that there existed a man with the name Egill Skalla-Grímsson
in the 10th century but also that we can believe in the Egill presented in Egils saga as a representation of that living Egill, apart from, of course, a few exaggerations
and idiosyncrasies such as his precocious ability to versify at the age of three,
which Sigurður Nordal attributes to the boasting of an old man (1933, xii).
Thus began the influential Íslenzk fornrit tradition wherein the introductions to each volume in the series should include a
discussion of fact and fiction in a given saga and an attempt to distinguish between
the historical and real part of the text, on the one hand, and the invented and fictional
part, on the other. Nowhere was this task undertaken in a more orderly fashion than
in Björn Sigfússon’s introduction to Ljósvetninga saga (xix–xli) in which his chapter “genealogies and historical knowledge” is followed by the chapter “fiction,”which he describes as things “sem getur ekki stuðzt við söguleg rök” [that cannot stand as historical evidence]. Björn also exemplifies the logical method used to distinguish between the two when
he states that the most reliable part of any saga must be the genealogies (xxx).
Sigurður Nordal and his colleagues admired art but did not like certain aspects of
the fictional such as the fantastical or paranormal, or what they considered exaggerated
or unrealistic. Thus they imagined the sagas as something betwixt and between history
and fiction or, in Sigurður Nordal’s mind, a union of critical historical thought
and a fine sense of art and entertainment (see e.g. Sigurður Nordal 1933, lxiii). He described this perfect harmony of art and science as gradually tending towards
historical fiction during the course of the 13th century before it was ruined in the
14th century, unwinding with the production of either dull annals or exaggerated romances,
clearly inferior to the work of Snorri Sturluson, the master of the harmonious sagas.
However, this invented category, that bears a close resemblance to contemporary historical
fiction, is determined by a modern distinction between the two forms as imagined opposites.
Thus the insistence on the literary value of the sagas was entwined with a consequent
and ever-growing rejection of the sagas as historiography (see Sveinbjörn Rafnsson 1988, 324), which was founded on a supposed equivalence between history and truth. The Íslenzk fornrit editors and most other Old Norse scholars of the period tended to demonstrate a lack
of understanding of history as a literary genre, and the debate over the historicity
of the sagas was hemmed in by the false opposition between the real and the artistic.
In 1974, Sveinbjörn Rafnsson controversially stated in his doctoral dissertation on
Landnámabók, conventionally regarded as a work of history rather than one of art, that it had
no value as an accurate source describing the settlement of Iceland (see also Sveinbjörn Rafnsson 1976, 232). This sparked strong reactions from various critics, including Jakob Benediktsson
(1974, 213) who had edited the text a few years earlier in the Íslenzk fornrit series and who felt that Sveinbjörn Rafnsson was too dismissive of the source value
of the text when it came to personal history. According to Jakob Benediktsson, denying
the source value of Landnámabók amounted to dismissing it as a “work of history.” In his own introduction to the
work (1968, cxxxiv), Jakob Benediktsson had followed the tradition of distinguishing between the “historical parts” of the text (genealogies and place names) and the “fictional parts” (such as the accounts of trolls and paranormal activity). Sveinbjörn Rafnsson, on
the other hand, refused to divide the text into the historical and the fictional.
The main difference between the two scholars lay, on the one hand, in their respective
beliefs regarding the sagas’ applicability, as Sveinbjörn Rafnsson felt that Landnámabók was an important source to many questions concerning the Icelandic society in which
it was composed, and other scholars have indeed since followed in his footsteps and
used the sagas to analyze systems rather than sources for information about individuals
(see e.g. Byock). On the other hand, the difference lies in their respective attitudes towards picking
and choosing. For Sveinbjörn Rafnsson there is no such thing as partial source validity,
and Landnámabók is simply too young to be an accurate source about the settlement, no matter how
credible some of the information that it provides may seem. He also objected to the
practice of 20th-century scholars to use their own judgment to decide what is likely
factual and what is improbable and then equating the likely factual with truth.
No recent scholar has professed a belief in Egill Skalla-Grímsson versifying at the
age of three—apparently the analogy of young Mozart fails to convince when juxtaposed
with the grim unintelligibility of skaldic poetry. However, Jón Helgason elegantly
summarized this kind of critical stance when discussing scholars who believed in the
account of Egill Skalla-Grímsson’s visit to King Eiríkr Blood-Axe in York. Jón Helgason
argued that since the whole story is based on the premise that magic drew Egill to
York, the scholars could either believe the whole story, including then the magic
of Queen Gunnhildr, who famously changes shapes to disturb Egill in composing the
poem that eventually saves his neck, or dismiss it in its entirety. In Jón Helgason’s
mind, scholars cannot simply pick and choose which events to believe in, as if it
is possible that “ef galdur er tekinn úr galdrasögu, verði afgangurinn sönn saga” [if magic is removed from a tale of magic, then what remains is a true story] (1969, 156).
Although this criticism has continued to be ignored by many scholars, scholarship
appears to be rapidly moving away from the picking and choosing or Íslenzk fornrit method, and 13th-century historiography is no longer regarded to be as reliable as
it once was when it comes to the history of the 9th, 10th, or 11th centuries. Scholars
will consequently have to study these sources as texts and as historiography instead
of focusing on and attempting to recover a lost “reality” that may lie behind them.
But what, then, happens to the representation of the real in the sagas? Many scholars
now seem to agree that in this context dated terms such as history, literature, reality,
fiction, and truth must either be abandoned or provided with a new sense (see e.g. Hermann and Mitchell 263). Throughout the 20th century it was, however, rare for scholars to consider the sagas
as history using modern (or post-modern) concepts of the genre.
One such attempt was made by Keld Gall Jørgensen (267–68) who sought to highlight the subjectivity of all truth, whether historical or fictional.
He was inspired by the recent emphasis on the literary value of history demonstrated
by those analysts of historical discourse who regarded history first and foremost
as narrative. Hayden White, for example, inspired by post-structuralism and noting
the gap between the philosophical discourse on reality and the work of traditional
historians that tended to ignore this discourse in their fieldwork, took a firm stance
against 19th-century empiricism. White referred to history as “the discourse of the real,” meaning that he regarded history not as reality but a representation of it, usually
taking on the form of a narrative. Thus history is separated from both events and
“reality” and is diagnosed as a part of language (which is also part of reality but a different
The medievalist Gabrielle Spiegel works in the same vein, although she identifies
her influences not as Barthes, Lacan, and White but rather Foucault, Geertz, and Turner.
In Spiegel’s estimation history is essentially a part of language and not reality,
and thus the opposition between history and literature is meaningless. As such, scholars
would be well served to focus more on historiography itself and its narrative devices
and less on any supposed inherent reality. Both White and Spiegel emphasize that any
study of history as a kind of discourse has to take into account the essential nature
of this discourse as being something different from overt fiction. The above-mentioned
framework set out by Sigurður Nordal where we can imagine a singular form that constitutes
a mystical union of history and fiction may not be helpful here since all history
is narrative and there is really no separation between history and literature.
If the sagas are regarded as historical texts that may or may not represent reality
accurately but are still much governed by the laws of narrative, any discussion of
their art will not lead to a negation of their intrinsic nature as works of history.
There is thus little sense in dividing the sagas into perceived historical and fictional
parts. They purport to be historical but those sagas relating events from the distant
past have little source value when it comes to factual information about bygone events.
Their source value lies more in what they reveal about their anonymous authors and
about the time and place in which they were written. Yet, this is not where their
overt engagement lies, and scholars who attempt to regard these narratives about the
past as allegories of the 13th- or 14th-century present, as has often been fashionable
in recent decades, may be taking too little note of a genuine engagement with the
If the emphasis is placed upon subjectivity, as Jørgensen does, instead of an absolute
truth that is in any case beyond human experience, it becomes evident that no “reality” can be gathered from historical sources like the sagas. Indeed all history must represent
the point of view of an author or authors, no matter how objective they intend or
pretend to be. In the case of the sagas of Icelanders, an added complication consists
in their distance from the events they purport to depict. However, even when there
are several contemporary accounts and a great deal of data, it does not necessarily
follow that “reality” can be gathered from the available sources, even when there is less distance between
the event and its accounts. Thus Jørgensen’s approach does not only mean that we must
doubt the source value of the sagas on account of their temporal distance from the
reality they purport to relate, but that we must call into question the perceived
objectivity of any kind of historical truth, even when the situation is quite different,
with a wealth of data and a closer proximity to the reported events.
If we take, for example, a relatively straightforward modern event such as the Kennedy
assassination, this problematic relationship between truth and data becomes equally
evident. This was a public event with many observers and an abundance of contemporary
data, and yet considerable ambiguity remains. Of the 104 earwitnesses in Dealey Plaza,
for example, who are on record having stated an opinion as to the direction from which
the shots were fired, 54 thought that all of the shots came from the direction of
the Texas School Book Depository, 33 from the grassy knoll or the triple underpass,
nine from a location entirely distinct from the knoll or the Depository, five heard
shots fired from two locations, and three from a direction consistent with both the
knoll and the Depository (McAdams).
Human experience is limited. Five people may sit in the same room but will only experience
the same reality up to some point. Human cognition immediately steps in and begins
framing reality, using thought and language and that strange attribute called memory,
which is our way of making sense of the world from within the recesses of the mind.
Human memory is essentially highly selective and thus the transformation from event
to narrative will always be framed by the devices of memory. First the event takes
place, but then it is experienced and after that interpreted and then memorized and
transformed into internal language before it is ever actually expressed, often presumably
orally before writing is ever employed.
When history is written, even more framing takes place. After hearing and remembering,
our subjective experience of the event is made into a narrative using human language,
and each of these actions inevitably transforms the reality of the event, first within
the reality of the mind and eventually into that of a given narrative. In the end,
even the matter of the number and direction of shots, as discrete a fact as is possible,
becomes a highly controversial point. Most events related in a saga, or other narratives,
are more complicated than determining who shot whom and with added complication comes
the added significance of the point of view of both sources and their authors.
In the case of the Kennedy assassination a commission (The President’s Commission
on the Assassination of President Kennedy, unofficially known as the Warren Commission)
was appointed to establish “truth” and, after interviewing more than 500 people, handed in an 889 page report on the
event. This process certainly established a truth, but not a truth that was convincing
to all. Yet the commission had the opportunities and the capacity to establish a “truth” that far exceed those available to any medieval historiographer. Of course, as any
conspiracy theorist will proclaim, the commission may also have had its own agenda,
but that must in some way, to varying degrees, be true about every historiographer.
The uncertainty shrouding this recent modern event is indeed enlightening in that
it shows that any discussion of the quest for the truth behind the sagas has to take
account of the very subjectivity of historical truths, even in the most favourable
of circumstances. When faced with the utter ambiguity of truth, the attempts made
by the editors of the Íslenzk fornrit series and other 20th-century scholars to separate fact from fiction in the sagas
thus seem even more desperate. How can they possibly hope to achieve this with the
limited knowledge at hand? What the Warren commission really presented was the result
of an investigation, a version of events that may or may not be accurate and that
will continue to convince some but certainly not all. Even less can medieval historians
be relied upon to bring us “the truth,” and perhaps this was something of which they were themselves well aware. They may
have been doing their best on this front, but, when narrating a past that was already
over 300 years old, there were certainly limitations as to what their best could possibly
be. Though possibly more disinterested than the Warren commission was, it would have
been hard to keep their own subjectivity from exercising an influence on the text.
As Nietzsche somewhat pessimistically declared half a century before the first volume
in the Íslenzk fornrit series appeared in print, history always stands in the service of the unhistorical.
It thus seems that we cannot entirely deny subjectivity in any historical narrative
regardless of how far removed it is from the events that it relates.
If we try to bring this insight to settlement-age Iceland and bridge the gap between
saga studies and the humanities at large, it may still be possible to believe that
individuals called Ingólfr and Skalla-Grímr, for example, lived there during the
late 9th century, but all the facts presented about them in the late medieval sources
are indeed the subjective image of a much younger age, a distillation of experience,
memory, interpretation, and narrative. These images are presumably a part of a long
narrative tradition that has transformed reality into a new reality above whatever
could have been found in a contemporary account. Skalla-Grímr may have had a son called
Egill, but the colourful character from the saga cannot be anything but a mixture
of what happened, what people experienced and remembered, and a long and volatile
tradition about this character, which was then amplified by the art of the anonymous
13th-century author or authors.
It is hard to say precisely what part of the Egill Skalla-Grímsson we can see in
the saga comes from the author and what was drawn from tradition, but one thing is
certain: a single “truth” cannot possibly be distilled from these sources. Modern
readers can believe in the existence of an Egill, but then they have to decide which
Egill. We never know whether this factual Egill, if he existed at all, was somewhat
like the character from the saga or quite different. Believing in Egill, like believing
in God, turns out to be no simple matter.
The most fascinating lesson to be drawn from the uneasy relationship that the study
of Old Norse texts has had with notions of truth, history, and fiction is how fascinated
humans, scholars, and amateurs alike are by what is considered real, accurate, and
truthful. They want the books that they read to be real, and they want to imagine
a reality behind the late medieval depiction of settlement and saga-age Iceland. Thus
critical historians have always faced an uphill task. All they have to offer is uncertainty
and doubt, so much more uncomfortable and less reassuring than legends that can be
believed in and regarded as “truth.”
The idea of fiction pursued by the Íslenzk fornrit editors (and by some modern scholars) may perhaps have been somewhat naïve as well.
The frequent references to a reality behind the text and the idea that a saga character
is not fictional if he or she is based on the reality of a 10th-century human is often
juxtaposed with the idea of fiction that is simply invented by an author out of the
blue. However, what is today called fiction is no less based on reality than that
which is called history. The rules are different but the writers or compilers of both
are composing texts that refer in some way to a reality. This applies also to the
medieval saga authors and it thus seems meaningless to categorize the matter in their
texts as either exclusively fictional, on the one hand, or real, on the other. The
biography of a saga character such as Egill Skalla-Grímsson is neither real nor fictional.
It is history as it was done in 13th-century Iceland, based on tradition with a great
deal of creative input by the historiographer. Trying to find “reality” in such a mixture of the traditional and the novel, with no contemporary or even
slightly older sources at hand to validate more than a tiny fragment of the information
provided, is a nearly impossible task. What we instead have is both tradition and
legend, and an interesting historical text.
A quarter of a century ago Sveinbjörn Rafnsson referred to the settlement age of Iceland
as desk fiction made into a view of history (1988, 319), and indeed the same can be said of the saga age of Iceland. Like King Arthur, it
is essentially legendary, a carefully constructed 13th- and 14th-century vision of
9th-, 10th-, and 11th-century Iceland. Just as no truth behind the legends can be
wrested from the textual sources of 6th-century Britain, we must likewise accept Ingólfr
and Egill and all the saga heroes of Iceland as textual representations of human beings
that had possibly lived during the 9th, 10th, and 11th centuries. The point is not
whether these people existed, though, since we do not have them in the sagas in any
case: we have characters and representations.
When it comes to the customs, ideologies, and social realities of the Icelandic society
of the 12th, 13th, and 14th centuries, the sagas have served as a mine of information
in recent years. When it comes to certain individuals, they yield a great deal of
traditional matter some of which may seem more plausible than the rest, but plausibility
is more a matter of belief than a useful marker to discover what is real. This may
not prove a great loss for textual criticism, as there remain compelling legends of
interesting characters from a 13th-century saga whose words and actions can still
be analyzed and studied.