Speech-WranglingShutting Up and Shutting Out the Oral Tradition in Some Icelandic SagasBrian McMahon

The ultimate origins of the Icelandic sagas are lost in the mists of time. The word saga (from Old Norse segir) means ‘that which is said or reported’; however, the written prose texts which describe themselves as sagas bear all the hallmarks of having been composed by literate authors and are now the only evidence that attests to a once apparently thriving oral storytelling milieu.11See Paul Bibire, ‘On Reading the Icelandic Sagas: Approaches to Old Icelandic Texts’, in West over Sea: Studies in Scandinavian Sea-Borne Expansion and Settlement before 1300, ed. by Beverley Ballin Smith, Simon Taylor, and Gareth Williams (Leiden: Brill, 2007), pp. 3–18 (p. 3). We do not know precisely when the Icelanders began to write their sagas down, but it is highly unlikely that the practice began in earnest before the middle of the eleventh century, and its development was certainly gradual.22Margaret Clunies Ross, The Cambridge Introduction to the Old Norse-Icelandic Saga (Cambridge: Cambridge University Press, 2010), p. 48. Since many sagas contain apparently accurate historical details inherited from the earliest settlers in the late ninth century, it follows that at least these snippets of information, embedded in narratives of indeterminate length, had been in circulation for some two hundred years before the literate sagamen began their task of composing — or, at least, redacting — written saga texts. We seldom know who these authors were, although the scribes of some manuscripts have been identified, but we do know that they ranked among the literate social and cultural elite in Iceland, primarily made up of clergy, lawyers, and landowning chieftains.33Pernille Hermann, ‘Literacy’, in The Routledge Research Companion to the Medieval Icelandic Sagas, ed. by Ármann Jakobsson and Sverrir Jakobsson (Abingdon: Routledge, 2017), pp. 34–47.

The oral tradition through which folk memory was given voice before the advent of literacy in Iceland was, by definition, comparatively open. Anyone who reported news was, in the literal sense of the word, telling a saga. Early writers thus had access to a diffuse nexus of stories from which to draw and shape the versions of the stories they would imprint upon the page. Stephen Mitchell imagines the authors of the great family sagas, for which we rarely have any witnesses dated earlier than the late thirteenth century, as each more closely resembling ‘a medieval Burns or Scott (or, perhaps more aptly, a medieval Paul Anderson or Michael Crichton)’ rather than a diligent antiquarian or folklorist determined to preserve the pure distillation of some ephemeral oral ur-saga.44Stephen A. Mitchell, ‘The Sagaman and Oral Literature: The Icelandic Traditions of Hjörleifr inn Kvensami and Geirmundr heljarskinn’, in Comparative Research on Oral Traditions: A Memorial for Milman Parry, ed. by John Miles Foley (Columbus, OH: Slavica, 1987), pp. 395–423 (p. 413). Tommy Danielsson has employed the metaphor ‘det muntliga havet’ (the oral sea) to express the fluid relationship between different oral iterations of the same stories which resist assuming a fixed form, since every recitation and repetition will differ from the last.55Tommy Danielsson, Sagorna om Norges kungar: Från Magnús góði till Magnús Erlingsson (Hedemora: Gidlunds Förlag, 2002). By contrast, as Ward Parks writes, ‘the written text could be defined as memory concretized […] fixed in durable form that frees it, apparently, from the effects of time’.66Ward Parks, ‘The Textualisation of Orality in Literary Criticism’, in Vox Intexta: Orality and Textuality in the Middle Ages, ed. by A. N. Doane and Carol Braun Pasternack (Madison: University of Wisconsin Press, 1991), pp. 46–61 (p. 58). The written sagas might therefore be conceived of as islands rising up out of the oral sea — discrete; with shorelines that erode just a little over time but retain their essential integrity; and solid rather than fluid, manifesting a particular version of the story, closed off from its original sources and influences, which remains in situ as the waters recede and the oceans drain away. There is every reason to suppose that the growth of the written sagas occurred in the context of a sustained oral tradition and that, at some stage, the two forms of any given story — the written and the oral — might have come into contact and, indeed, competition with one another.77A common phenomenon in medieval Europe, discussed at length in D. H. Green, Medieval Listening and Reading (Cambridge: Cambridge University Press, 1994). Yet they were distinct in modal terms: the liquid oral tradition was an open one; the process of inscribing texts on parchment sealed them off, to some extent, from further innovation. Where different or contradictory written variants emerged, their differences and contradictions could no longer be elided through dialogue and exchange between living storytellers passing on mutable stories; the writing down of sagas was, in this respect, an act of closure.

Given these circumstances, it is quite remarkable that there is such uniformity within the surviving corpus of the written sagas. Certainly, there are differences of expression which distinguish one redaction from the next — sometimes amounting to the inclusion or omission of whole episodes — and certainly, as Carol Clover has written, ‘the sagas share characters, dovetail matter, and refer and defer to one another in a way that suggests that they were not conceived as self-contained wholes but as interrelated or interdependent members of a larger undertaking’.88Carol Clover, The Medieval Saga (Ithaca: Cornell University Press, 1982), p. 41. Yet for all that, there is just one surviving Njáls saga, just one Grettis saga, a single Laxdœla saga. These exist in variant versions, but each clearly descends from a common source — although that source itself may once have been compiled from different oral influences. As Gísli Sigurðsson has argued,

it is not unlikely that the plot and subject-matter of the sagas was derived from a living tradition of oral story-telling, where it was moulded by performers interacting with their audiences until it eventually received its fixed form in a written saga designed to be read99Gísli Sigurðsson, ‘Another Audience — Another Saga: How Can We Best Explain Different Accounts in Vatnsdœla saga and Finnboga saga ramma of the Same Events?’, in Text und Zeittiefe, ed. by Hildegard L. C. Tristram (Tübingen: Narr, 1994), pp. 359–76 (p. 375).

Of course, the promulgation of the written saga does not exclude the likelihood that various versions of the same narrative continued to circulate and evolve orally. The proliferation of manuscripts attests to the enduring popularity of these stories, and they frequently invoke one another, but the lack of competing traditions — different sagas telling the same story, rather than variant versions of the same saga — is certainly striking and significantly at odds with the mythological and heroic verse literature recorded during the same period. Two obvious solutions present themselves: either Icelandic institutional memory really was so exceptional that the same version of the same story persisted throughout the country for more than two centuries; or at some stage an editorial process took place through which unauthorized sagas were excluded from the written corpus in favour of a single, preferred, ‘best-text’ iteration of each story. Such a process, if it occurred, would be difficult to examine, since by definition it would have involved the expunging of competing sagas dealing with the same individual, area, or episode, but it may be possible to deduce its likelihood from the evidence of cultural attitudes which the surviving sagas provide. Editing the corpus in this way — redacting the oral tradition into authorized written versions of the sagas — would represent a process of closing off access to divergent versions in favour of a single approved iteration.

In addition to Tommy Danielsson’s metaphorical ‘oral sea’, I would like to propose an analogy with the modern concept of copyright. The oral tradition must have been comparatively open to revision from a range of sources — susceptible to changes emerging in the narrative, focus, and ductus of the sagas — but by constraining or ‘enclosing’ the story within a manuscript and fixing it within the limits of the page, its early editors began to restrict these possibilities, gradually confining the saga within set narratological, orthographical, and codicological boundaries. The act of inscribing a saga on parchment required literacy, scribal expertise, and the expenditure of resources. As such, the process of writing down a particular redaction of the narrative conferred a certain status upon it. Writing the saga down therefore represented a challenge to alternative versions of the story then in circulation. The act of writing also represented the imposition of limitations on the scope of the story — a beginning and end within which this discrete saga took place. An oral storyteller, working from memory rather than from a manuscript, might improvise and innovate in the course of a recitation, modulating his or her performance in response to live audience feedback, but the writing down of the saga restricts the private reader from taking similar liberties. Emendation would be possible, insofar as there was space on the page, but the skeleton of the text would now be fixed. The opening and closing of the book controlled access to the written saga, and therefore access to the book was necessary to access this high-status redaction of the traditional story. Additionally, the writing down of sagas restricted access to those who could read — in other words, the literate elite.

Whereas the practice of oral storytelling was potentially open to all sectors of society, written sagas were available only to a closed community of readers — those who both enjoyed access to the manuscripts and possessed the literate skills necessary to glean their content from them. This community might be opened up to a wider audience through the reading aloud of sagas from a manuscript, but such a process should still be considered less ‘open’ than the preliterate oral tradition, since it could only take place subject to the availability of a manuscript and the presence of a suitably qualified (i.e. literate) reader. At the very least, the production of written saga texts introduced a three-tier system for the reception of these stories: they were either spoken aloud from memory, read aloud from a manuscript, or read privately by a sufficiently competent individual, who was also therefore exposed to paratextual material which might not necessarily be communicated through an oral performance.1010D. H. Green refers to texts designed with an eye to public as well as private transmission as an ‘intermediate mode’ of storytelling (‘Orality and Reading: The State of Research in Medieval Studies’, Speculum, 65.2 (April 1990), pp. 267–80). This emerging distinction between the written word and oral culture, potentially freighted with hierarchical associations for each means of reception, was doubtless in Oddr Snorrason’s mind when he cautioned readers of his Óláfs saga Tryggvasonar to prefer his written redaction of the saga over other competing stories which they might have heard: ‘Ok betra er slict með gamni at heyra en stivp meðra saugvr, er hiarðar sveinar segia, er enge viet hvart satt er’ (And it is better to listen to such [tales] with enjoyment than to stepmothers’ stories, which shepherd-boys tell, which nobody knows the truth of).1111Saga Óláfs Tryggvasonar, ed. by Finnur Jónsson (Copenhagen: Gads Forlag, 1932), p. 2. Translations, except where otherwise stated, are my own. It is significant that Oddr appears to criticise both the source of these ‘stivp meðra saugvr’ (stepmothers’ stories) and their mode of transmission — spoken rather than read. Both characteristics appear to indicate their low status in this author’s mind, a fact perhaps reflective of his dual profession as a Benedictine monk and a scribe. In both capacities, Oddr would have been a natural champion of Scripture and the written word, predisposing him to look sceptically at oral tradition as a potential vehicle for pagan (or, at least, unorthodox) wisdom and practices.1212On Oddr’s prologue, see Judy Quinn, ‘From Orality to Literacy in Medieval Iceland’, in Old Icelandic Literature and Society, ed. Margaret Clunies Ross (Cambridge: Cambridge University Press, 2000), pp. 30–60 (pp. 38–40). His subject, Óláfr Tryggvason, actively fought paganism in Scandinavia, and Oddr’s awareness of the novelty of monastic life in Iceland (his monastery at Þingeyrar being the first to be founded there in 1133) would have provided him with an incentive to be a champion of written, authorized, Christian histories over and above competing oral iterations of the same stories. In this respect, his warning reflects a wider medieval tension between pre-Christian oral narratives and their post-conversion written descendants.

The high medieval Icelandic elite had ample motive for wanting to control and ‘authorize’ the writing down of the sagas. As Theodore Andersson writes, ‘the content of the stories was no doubt agreed on by many people, but the selection and ordering of the stories was left to the individual teller or writer who shaped them’.1313Theodore M. Andersson, The Growth of the Medieval Icelandic Sagas (Ithaca: Cornell University Press, 2006), p. 19. This ‘shaping’ amounts to the imposition of control over the sagas, and this is especially pertinent in the case of the family sagas (Íslendingasǫgur) which, along with Landnamabók (a medieval record of the early settlement of Iceland, possibly first compiled in the late eleventh century) and Íslendingabók (an early twelfth-century history of Iceland by Ári Þorgilsson), comprise the story of the founding of Iceland, Europe’s only medieval commonwealth, by a proud and independent people whose descendants had a vested interest in their commonly agreed content. History is written by the victors — that is, the ruling elite, and it would be in their interests to establish a widely circulated and accepted basis for their present high status. As Kirsten Hastrup has written,

In Icelandic, saga means both story and history. It is literally what is ‘said’ about previous events, periods, or people. Telling makes history. The Icelandic sagas are stories of different historical veracity, but the point is that in the concept of saga, story and history are one.1414Kirsten Hastrup, A Place Apart: An Anthropological Study of the Icelandic World (Oxford: Clarendon Press, 1998), p. 23.

Paul Bibire goes still further in claiming that any attempt to impose a modern distinction between ‘literature’ and ‘history’ is ‘irrelevant to the study of Norse’.1515Bibire, ‘On Reading the Icelandic Sagas’, p. 15. This disjunction between modern and medieval attitudes towards history naturally extends beyond the Norse-speaking world. The Latin term historia is similarly multivalent, meaning both ‘history’ and ‘narrative’; however, the Old Norse word saga is still more complicated because of the explicit allusion to speech which it contains. The adoption of this term for written texts suggests a certain conservative desire either to retain the impression of spoken history or to appropriate and control it by imprinting it on vellum. Thus, while the reluctance to distinguish between fiction, legend, and history which frustrates modern historians reflects a widespread medieval European mindset, the Old Norse sagas conflate not only fact with fiction but also the spoken word with the written. Most of the early surviving redactions of these sagas were written in the Sturlung Age (1220–64), when internecine warfare threatened the stability of Iceland and imperilled its independence.1616Iceland became a vassal state under the Norwegian Crown in 1262. During and after this period there were strong reasons to compose a record — however embellished — of the country’s earlier glory. In Hastrup’s words, ‘by stressing the unity of people, history, and language an ideology of Icelandicness [was] created’.1717Hastrup, A Place Apart, p. 90. Additionally, we know that many sagas were commissioned, with the name of their commissioner being associated with them rather than that of the author, compiler, redactor, or scribe. In this sense they bore the authority of he who caused them to be made, and he — who was, of necessity, wealthy — would likely resist the persistence of alternative forms of the same saga which undermined or contested his proxy composition.

Insofar as surviving texts of the sagas do vary, the variation tends to be most extreme in the prologue or epilogue appended to the text in certain redactions. Here we occasionally encounter the voice of the scribe — or possibly the author; it is always difficult to distinguish between the two — emerging from behind the mask of studied anonymity which he otherwise wears throughout.1818Such narrative interventions are more common in kings’ sagas, but they can be found in some redactions of Íslendingasǫgur, including the epilogues to Droplaugarsona saga (see below) and Bolla þáttr Bollasonar (which describes how many accounts of Bolli’s journey are in circulation). For a thorough discussion of narratology in the sagas, see Heather O’Donoghue, Narrative in the Icelandic Family Saga: Meanings of Time in Old Norse Literature (London: Bloomsbury Academic, 2021). These fragmentary contributions from different periods and iterations in a saga’s development provide a rare glimpse into the agendas of those involved in its transmission, and are consequently a rich source of information about how authors, scribes, and compilers conceived of their respective roles in this process. Differentiating between these interacting voices is rarely straightforward, but the prolegomena which do survive deserve serious attention because they provide snatches of the discourse which took place between the saga texts and their medieval audiences, and potentially help to illuminate the designs of those who commissioned them in their written forms. One obvious example of an epilogue acting as a critical commentary on a saga text occurs at the end of Oddr’s Óláfs saga Tryggvasonar. As Andersson points out, this passage ‘stakes an Icelandic literary claim: in effect, it copyrights the biography of Olaf Tryggvason’, appropriating this famous king of Norway as an honorary Icelander.1919Andersson, Growth of the Medieval Icelandic Sagas, p. 204. This distinction would have been significant for many Icelanders, who continued to regard their commonwealth as having been in tension with the Kingdom of Norway ever since the earliest Scandinavian settlers in Iceland broke with King Haraldr hárfagri in the ninth century. Despite often relying on the patronage of the Norwegian kings, Icelandic saga heroes are typically proud of their fledgling commonwealth’s independence, and therefore any Icelandic history of a king believed to be descended from Haraldr — as Óláfr Tryggvason was — would be sensitive to the competing claims of oral biographies circulating between mainland Scandinavia and Iceland.2020See further Theodore M. Andersson, ‘The First Icelandic King’s Saga: Oddr Snorrason’s “Óláfs saga Tryggvasonar” or “The Oldest Saga of Saint Olaf”?’, Journal of English and Germanic Philology, 103.2 (2004), pp. 139–55.

If Oddr’s version of the saga was to achieve the status of being the definitive record, he would need to authenticate it in some way and indemnify it against future competition from competing versions, both oral and, potentially, written. He appears to attempt this in an additional chapter appended to the saga in one manuscript, AM 310 4^to (c. 1250–75), which contains the following passage:2121Its source, though, may have been his fellow monk Gunnlaugr Leifsson. See further Bjarni Aðalbjarnarson, Om de norske kongers sagaer (Oslo: Dybwad, 1937), pp. 85–86, for the case against Oddr having shown his text to Gizurr Hallsson. This argument is persuasively refuted in the introduction to Andersson’s more recent translation: The Saga of Olaf Tryggvason, trans. by Theodore M. Andersson (New York: De Gruyter, 2003), pp. 3–4.

Þessa sogu sagþi mer Asgrimr abboti Uestliða s. Biarni prestr Bergþors s. Gellir Þorgils. s. Herdis Daða dottir. Þorgerðr Þorsteins. d. Inguðr Arnors. d. Þessir menn kendu mer sua sagu Olafs konungs T. s. sem nu er sogð. Ec synda oc bokina. Gitsure Hallz s. oc retta ec hana eptir hans raðe.2222Saga Óláfs Tryggvasonar, p. 247.

(I was told this story by Abbot Ásgrímr Vestliðason, the priest Bjarni Bergþórsson, Gellir Þorgilsson, Herdís Daðadóttir, Þórgerðr Þorsteinsdóttir, [and] Inguðr Árnórsdóttir. These people instructed me in the saga of King Óláfr Tryggvason as it is now told. I showed the book to Gízurr Hallsson and corrected it with his counsel.)

This careful referencing of multiple sources and deference to an established authority for correction represents a marked attempt to set the text apart from its oral antecedents which, so far as we can tell, deliberately avoided association with particular sources, being framed rather as a continuance of unbroken (and thus relatively ‘open’) oral discourse — a convention also to be found in Ári Þorgilsson’s Íslendingabók (c. 1122–33). That phrase ‘sem nu er sogð’ (as it is now told) is especially instructive, since it implies a conscious effort to distinguish the present iteration from any competing — allegedly spurious — versions. The written text is not merely ‘the saga’ but ‘the saga as told here’ — the authorized redaction. This effect is substantially amplified by the litany of authorities to which the redactor refers. Rather than offering himself as a reliable source per se, the author of this passage cites the names and credentials of prominent Icelanders with a reputation for wisdom (reputations in turn burnished in many of the so-called ‘contemporary’ sagas, or samtíðarsǫgur). The rhetorical effect is to suggest the conferring of authenticity by consensus. The scribe depicts a process by which he first ‘opens’ his text for comments and contributions from a wider polity of knowledgeable experts before emphatically ‘closing’ it by inscribing it on parchment as an approved testamentary record.2323For a discussion of similar appeals to authority in relation to Old Norse legal texts, see Stefan Brink, ‘Minnunga mæn: The Usage of Old Knowledgeable Men in Legal Cases’, in Minni and Muninn: Memory in Medieval Nordic Culture, ed. by Pernille Hermann, Stephen A. Mitchell, and Agnes S. Arnórsdóttir (Turnhout: Brepols, 2014), pp. 197–210.

This insight may also shed light on the famous coda to Droplaugarsona saga, which reads: ‘Þorvaldr átti son, er Ingjaldr hét. Hans sonr hét Þorvaldr, er sagði sǫgu þessa’ (Þorvaldr had a son, and he was called Ingjaldr. His son was called Þorvaldr, who told this story).2424Droplaugarsona saga, ed. by Jón Jóhannesson, Íslenzk Fornrit, 11 (Reykjavik: Hið Íslenzka Fornritafélag, 1950), p. 180. Tempting as it has always been to consider this a generically typical third-person reference to the author of the extant saga, it is surely more likely that the informant, Þorvaldr, is named as an authenticating voice only, and not, as Peter Hallberg thought, ‘enough to prove that the family sagas were not in principle regarded as anonymous’.2525Peter Hallberg, ‘The Syncretic Saga Mind: A Discussion of a New Approach to the Icelandic Sagas’, Mediaeval Scandinavia, 7 (1974), pp. 102–17. For an alternative view, see Ralph O’Connor, ‘History or Fiction? Truth-Claims and Defensive Narrators in Icelandic Romance-Sagas’, Mediaeval Scandinavia, 15 (2005), pp. 101–69 (p. 114). Pragmatically, identifying an author would do nothing to authenticate the saga — it may, in fact, have had the opposite effect of suggesting literary or editorial innovation rather than faithful historical chronicling — whereas naming the saga’s source preserves a sense of proximity to the action it relates. Þorvaldr may indeed have sagði (told) the news, but it was samansetta (assembled) by others from the raw material of history and tradition. The voice of ancient sources speaks louder for the saga’s authenticity than that of even the most erudite later author. A similar attempt to associate a saga with a known authority (who is most unlikely to have authored it per se) can be found in one redaction of Gunnlaugs saga ormstungu, which claims to follow the version of the story given by Ári Þorgilsson, prefacing the saga text with a single sentence that contains no fewer than three references to his renowned wisdom, which it thereby seeks to co-opt and associate with the text that follows.2626The manuscript in question is Holm. Perg. 18 4to in the Royal Library, Stockholm. This gambit is presumably intended to pre-emptively close down any opportunity for dissent.

Implicit in these passages is the desire among the literate classes to create a canon of accepted story variants — an agenda frequently discernible in the text of later written sagas such as Grettis saga Ásmundarsonar, which at one point attempts to resolve a perceived disparity between divergent reports about its hero through an appeal to popular consensus:

Grettir var jafnan með Birni, ok reyndu þeir margan frœknleik, ok vísar svá til í sǫgu Bjarnar, at þeir kallaðisk jafnir at íþróttum. En þat er flestra manna ætlan, at Grettir hafi sterkastr verit á landinu, síðan þeir Ormr Stórólfsson ok Þórálfr Skólmsson lǫgðu af aflraunir.2727Grettis saga Ásmundarsonar, ed. by Guðni Jónsson, Íslenzk Fornrit, 7 (Reykjavik: Hið Íslenzka Fornritafélag, 1936), p. 187.

(Grettir was staying with Bjarni, and they tried many bouts, and it is said in Bjarnar saga that they were called equal at sports. But it is most people’s belief that Grettir was the strongest man who lived in the country since Ormr Stórólfsson and Þórálfr Skólmsson ended their strength-contests.)

By first acknowledging and then subsequently contradicting the earlier account, using popular opinion as authenticating proof, this saga seeks to establish a definitive version of events. Saga authors and scribes, at least by the fourteenth century, were clearly confident of the need both to acknowledge and to seek to supersede alternative narratives which recounted the same happenings, often offering a pre-emptive riposte to readers or listeners who might dispute their interpretation. Where a writer found himself not inclined or not able to proffer a definitive account, he would make reference to another saga and cede to it the greater authority concerning a particular subject.

Although the Íslendingasǫgur are famously circumspect concerning their redactors’ motivations, they do not avoid the topic altogether. Grettis saga is particularly distinguished by the attention it pays to the purpose of saga-telling, remarking after an account of the Battle of Havsfjord (c. 885) which occurs early on that ‘koma hér ok flestar sǫgur við, því at frá þeim er jafnan flest sagt, er sagan er helzt frá gǫr’ (the majority of sagas refer to it [the battle], because it is such matters that sagas usually refer to).2828Grettis saga Ásmundarsonar, p. 5. Whether the writer of these words had in mind only written sagas or written and oral sagas circulating simultaneously remains uncertain; what is noteworthy is the assertion that the saga is not an entirely open form but rather exists as a vehicle for certain kinds of material, of which this battle, part of the foundational narrative of Iceland, is an example. The author of this passage had strong evidence to support his claim, moreover, since a substantial number of Íslendingasǫgur do indeed begin with genealogies — often of Norwegian kings rather than Icelanders — followed immediately by a synopsis of the settlement. The term ‘Saga Age’, used by modern scholars as a device for distinguishing between Íslendingasǫgur and other genres of saga literature, indicates an enclosed period of time beginning shortly before the settlement (frequently dealt with in the prologue, even if it has little direct bearing on the nominal subject of the saga) and concluding with the conversion to Christianity (often supplemented by epilogues which assert the Christian credentials of saga protagonists, such as when Guðrún becomes a nun at the end of Laxdœla saga). These sagas are not simply a record of ‘what is said’, but rather of what is thought to be important by those chronicling this two-hundred-and-fifty-year period. The Íslendingasǫgur collectively close off this period from the present, confining pre-Saga Age genealogies to the prologue and most post-conversion concerns to the epilogue.

These observations help to account for the cursory tone often adopted at the end of Íslendingasǫgur, and it is noteworthy that these sparse epilogues afford virtually the only opportunity for the narrator, scribe, or author of the saga to address the reader directly. This is the case in Njáls saga, the longest and greatest in scope of the Íslendingasǫgur, which concludes with the words ‘Ok lýk ek þar Brennu-Njáls sǫgu’ (And thus I end Burnt Njáll’s saga),2929Brennu-Njáls saga, ed. by Einar Ólafur Sveinsson, Íslenzk Fornrit, 12 (Reykjavik: Hið Íslenzka Fornritafélag, 1954), p. 464. and much the same formula appears at the end of Sneglu-Halla þáttr: ‘Lýk ek þar sǫgu frá Snegu-Halla’ (And so I conclude the story of Sarcastic Halli).3030Sneglu-Halla þáttr, in Eyfirðinga sǫgur, ed. by Jónas Kristjánsson, Íslenzk Fornrit, 9 (Reykjavik: Hið Íslenzka Fornritafélag, 1956), pp. 261–95 (p. 295). A third example may be found at the end of Finnboga saga. In both cases the formulation is striking, since it represents the reader’s first and only direct encounter with the narrator of the story referring to himself in the first person. In this sense a direct encounter takes place only at the last possible moment, and the effect is akin to the removal of a mask or, perhaps, the laying aside of a manuscript from which the reciter of the saga has been reading so as to enable direct eye-contact with the audience. In the context of an oral recitation, this device might serve a number of purposes: to help ease the transition from the storyworld of the saga, closed off in historical time, to the present day; or to differentiate between the scribe and the reader. We might, for instance, consider the likelihood that these final sentences were not intended to be read aloud, but rather as a private remark for the eye of the literate reader, rather than the ear of his audience. They might communicate the subtext that this is a particular redaction of the saga — one compiled and controlled by the figure who identifies himself as ‘ek’ (I) in the closing lines, effectively signing off his authorized version of the story.3131Green, ‘Orality and Reading’, p. 277. See also Else Mundal, ‘How Did the Arrival of Writing Influence Old Norse Oral Culture?’, in Along the Oral–Written Continuum, ed. by Slavica Ranković, Leidulf Melve, and Else Mundal (Turnhout: Brepols, 2010), pp. 163–81. For a wider discussion of medieval attitudes towards orality and textuality, see M. T. Clanchy, From Memory to the Written Record, 3rd edn (Chichester: Wiley-Blackwell, 2013), esp. p. 298. The act of reading the saga aloud would represent an opening up of its contents to a wider audience, yet certain aspects of what appeared on the page might remain obscure to them, intended for the eye of the reader rather than the ear of the audience.

A contrasting, though similarly brief epilogue concludes Þórðar saga hreðu, and reads a little like a disclaimer, perhaps intended for the literate reader in the first instance and then, at his discretion, for members of a wider audience. This is the remark that ‘Þórðr hreða varð sóttdauðr. Höfum vér ekki fleira heyrt með sannleik af honum sagt’ (Þórðr the Menace died in his bed. We have not heard any more true facts about him).3232Þórðar saga hreðu, in Kjalnesinga saga: Jökuls þáttr Búasonar, Víglundar saga, Króka-refs saga, Þórðar saga hreðu, Finnboga saga, Gunnars saga keldugnúpsfífls, ed. by Jóhannes Halldórsson, Íslenzk Fornrit, 14 (Reykjavik: Hið Íslenzka Fornritafélag, 1959), pp. 161–226 (p. 226). Two observations can be made here: first, the qualification that the saga’s compiler knows of no more true facts recalls the language used in the prologue to Oddr’s Óláfs saga Tryggvasonar, and appears to suggest the continuing circulation of false reports which require shutting up and shutting out of the authorized record. Second, the use of the first-person plural pronoun distinguishes these closing comments from those examined above, since it implies a kind of corporate authorship — as though this redaction of the saga was the work of a committee, perhaps comprised of learned men like those listed by Oddr in his prologue or referred to in the closing lines of Droplaugarsona saga. This impression might again be intended for the eye of the private reader rather than the ear of an audience, but in any event it represents another attempt to refine — and thereby close down — the narrative of Þórðr’s life. Of course, an alternative reading of this coda might be to interpret it as an invitation; were the text of this epilogue to be read aloud, it is possible that a historically minded audience might wish to contribute ‘true’ stories from their own additional store of knowledge. What appears to be an act of closure might, if the literate reciter chose to read the whole passage aloud, prove rather an opening up of the storytelling ritual to accommodate a reciprocal exchange of knowledge or tradition about the life of Þórðr. Any claim to authenticity for the ‘true facts’ exchanged in this way would surely rely on the reputation of those who contributed them, with those participants known to be gifted with long memories, and perhaps those descended from Þórðr, likely to have been credited with special wisdom — perhaps exceeding even that codified in the book. It might be helpful to regard the process of writing the saga down as one of closure, and the reading of the saga aloud as one of opening up.3333Stephen M. Tranter, ‘Reoralization: Written Influence, Oral Formulation’, in Text und Zeittiefe, ed. by Tristram, pp. 45–54.

Alternatively, the use of the first-person plural pronoun may reflect an attempt to imitate or pay homage to the oral tradition through which the narrative is understood to have passed before reaching this fixed, static form. This kind of fictional orality, defined by Almut Suerbaum and Manuele Gragnolati as ‘the creation of a spoken, collective voice evoking poetic presence, but doing so by means of a consciously literate and literary written text’, is common to various medieval texts and genres, from the Nibelungenlied to the opening lines of Beowulf to numerous romances.3434Almut Suerbaum, in collaboration with Manuele Gragnolati, ‘Medieval Culture “betwixt and between”: An Introduction’, in Aspects of the Performative in Medieval Culture, ed. by Manuele Gragnolati and Almut Suerbaum (Berlin: De Gruyter, 2010), pp. 1–12 (p. 1). For a discussion of how the term ‘fictional orality’ can be applied in Old Norse contexts, see Stephen Mitchell, ‘Memory, Mediality, and the “Performative Turn”: Recontextualizing Remembering in Medieval Scandinavia’, Scandinavian Studies, 85.3 (2013), pp. 282–305. The use of the first-person pronoun remains strikingly uncommon in the Íslendingasǫgur corpus, however, and whatever its intended effect, the fact that its rare occurrences are almost entirely limited to prologues and epilogues reinforces the impression that the relationship between the narrative voice and the reader or audience is understood to shift here, at the limits of the text, a natural boundary between closed (i.e. formalized and monodirectional) and open discourse.3535Slavika Ranković, ‘The Performative Non-Canonicity of the Canonical: Íslendingasǫgur and their Traditional Referentiality’, in The Performance of Christian and Pagan Storyworlds, ed. by Lars Boje Mortensen, Tuomas M. S. Lehtonen, and Alexandre Bergholm (Turnhout: Brepols, 2013), pp. 247–72.

Similar passages which appear to contain the subtextual invitation to share knowledge occur in other sagas and þættir (‘short sagas’ or ‘fragmentary saga episodes’), suggesting that the storytelling community was not so closed as is sometimes thought, nor exclusively made up of literate Icelanders. For example, in Þorleifs þáttr jarlsskálds the narrator remarks: ‘ok gengr af honum [Hallbjǫrn] mikil saga bæði hér á landi ok útlendis, þó at hon sé hér eigi rituð’ (and there is a saga about him [Hallbjǫrn] that is well known here in Iceland and abroad, though it is not written here).3636Þorleifs þáttr jarlsskálds, in Eyfirðinga sǫgur, ed. by Jónas Kristjánsson, pp. 213–29 (p. 229). Any encounter with the written þáttr, whether as a private reader or member of the audience, must lead one to wonder about this story and seek to supply it from one’s external knowledge of these persons and events where possible. In this sense, the manuscript containing the þáttr represents as much a prompt book as a complete and enclosed narrative; the story of Þorleifr was written down in a kind of authorized redaction, but it continued to allude explicitly to supplementary material which was perhaps only available in the oral tradition at the time of writing or not immediately available to the redactor in a form which he considered to be authoritative.

Despite the impression either of anonymity or corporate authorship which the Íslendingasǫgur present, their narrators are not always hesitant about asserting their editorial powers and anticipating certain objections which their readers might raise. Eyrbyggja saga, for instance, announces itself as the ‘saga of the people of Eyri’, but early on the narrative voice remarks that ‘þarf hér ekki at segja frá þeira manna landnámum, er eigi koma við þessa sǫgu’ (there is no need to speak here about the settlements belonging to people who do not come into our story).3737Eyrbyggja saga, ed. by Einar Ólafur Sveinsson and Matthías Þórðarson, Íslenzk Fornrit, 4 (Reykjavik: Hið Íslenzka Fornritafélag, 1935), p. 11. Taken together with a similar remark from Grettis saga — ‘Mart bar til tíðenda um sameign þeira byskups ok Norðlendinga, þat er ekki kemr við þessa sǫgu’ (There are many stories about the exchanges between the bishop’s men and the men in the north, but these are not part of this saga)3838Grettis saga Ásmundarsonar, p. 35. — this has again the look of a disclaimer, reflecting the writer’s need to account for the decisions made in promulgating this particular redaction of the story. It is noteworthy that these statements do not pronounce on the importance of the redacted material, merely on its relevance to the narrative, or the extent to which it is suitable content for a text in the saga genre. Comments of this kind, common throughout the corpus, contribute to our impression of a collective endeavour towards dividing up the ‘oral sea’ and imposing static order upon it. This process might be termed ‘canonization’ and is, in any event, an act of enclosure, separating one saga — one set of incidents — from the next.

One of the enduring curiosities concerning the Íslendingasǫgur — and Old Norse sagas more generally — is that they should have been written in the vernacular. If the agenda of those who committed them to parchment was straightforwardly to generate a high-status written record of early Icelandic history, perhaps one thought to be of interest to the peoples of Scandinavia and wider Europe, then writing in Latin would have been the obvious choice. In any event, we might reasonably expect to find a mixture of languages, as we do a mixture of prose and verse, but in fact Latin passages — even Latin rubrics — are remarkably scarce. There is every reason to suppose at least some Latin literacy among most of the scribes responsible for the sagas, and therefore the fact of their being written in Old Norse indicates a specific and deliberate preference. One suggestive insight is offered by the anonymous author of the First Grammatical Treatise, a work dated to the mid-twelfth century, or early period of vernacular saga writing:3939Einar Haugen, ‘First Grammatical Treatise: The Earliest Germanic Phonology’, Language, 26.4 (1950), pp. 4–64 (p. 6).

J fleſtvm londvm ſetia menn a bækr annat tveggia þann froðleik er þar innan landz hefir giorz ęða þann annan er minniſamligaztr þikkir þo at annarſ ſdaða[r hafi] helldr giorz ęða lǫg ſin ſetia menn a bækr hverr þioð a ſína tvngv.4040Text and translation from The First Grammatical Treatise: Introduction, Text, Notes, Translation, Vocabulary, Facsimiles, ed. by Hreinn Benediktsson (Reykjavik: Institute of Nordic Linguistics, 1972), pp. 206–07.

(In most countries men record in books either the [historical] lore [relating to events] that have come to pass in that country, or any other [lore] that seems most memorable, even though it [relates to events that] have taken place elsewhere, or men commit their laws to writing, each nation in its own tongue.)

The implied connection between law and history suggests a common interest in maintaining records of both, while the mention of nations keeping these records ‘a ſína tvngv’ (in their own tongue) suggests a closed linguistic community. The situation is made slightly more complicated by the fact that modern linguistic distinctions were not necessarily recognized in the Middle Ages. The witness of Gunnlaugs saga ormstungu attests that the saga writer believed the language of the Anglo-Saxons and the tenth-century Icelanders to have been at least contiguous if not actually identical.4141For the relevant passage, see Gunnlaugs saga ormstungu, in Borgfirðinga sögur: Hœnsa-Þóris saga, Gunnlaugs saga ormstungu, Bjarnar saga Hítdœlakappa, Heiðarvíga saga, Gísls þáttr Illugasonar, ed. by Sigurður Nordal and Guðni Jónsson, Íslenzk Fornrit, 3 (Reykjavik: Hið Íslenzka Fornritafélag, 1938), pp. 40–108 (pp. 70–71). If this view was widely shared, it would suggest that differentiating between groups and cultures on the basis of language was less straightforward than the modern designation of the language used in in saga writing as ‘Old Norse’ (or, more specifically, ‘Old West Norse’ or ‘Old Norse-Icelandic’) initially implies. Nonetheless, the important point remains that the language is not Latin — the lingua franca of the elite — but a vernacular tongue particular to the descendants of those about whom the sagas were written. While immediate access was therefore restricted to those who could read, the oral recitation of a saga from a manuscript would have been widely understood because of the choice to record it in the vernacular tongue. The editorial control exercised by those who wrote sagas down was not wholly intended as an act of foreclosure, but rather of control and curation for a wide (though predominantly Icelandic) audience.

Scribes such as Oddr Snorrason had access to Latin texts and, indeed, often wrote their own compositions in Latin (the surviving Old Norse translations of his Óláfs saga Tryggvasonar appear to derive from a Latin original). Since Latin was the language of the Church and the universal language of European scholarship, the determination to write in the vernacular always represented a deliberate and particular choice. Many medieval Old Norse texts written in Iceland were translated from Latin exemplars, but the instinct to compose in the vernacular seems to have been unusually strong when compared to continental European cultures. Margaret Clunies Ross points out that the act of translating high-status texts such as saints’ lives into the vernacular likely had the effect of elevating the vernacular as a suitable language for expressing high-status ideas.4242Clunies Ross, Cambridge Introduction to the Old Norse-Icelandic Saga, p. 47. While Latin was plainly thought suitable for many kinds of texts, including some sagas, the Íslendingasǫgur were invariably written in Old Norse, this being the language of Iceland, and were therefore most immediately accessible to the descendants of their storied protagonists — a quasi-closed linguistic community around which the notion of a nation, independent from the Scandinavian mainland and, indeed, the European continent, was being formed.

Íslendingasǫgur are, without exception, anonymous.4343Notwithstanding Sigurður Nordal’s spirited attempt to demonstrate that Snorri Sturluson was the author of Egils saga, compelling proof has yet to be produced. See the introduction to Egils saga Skalla-Grímssonar, ed. by Sigurður Nordal, Íslenzk Fornrit, 2 (Reykjavik: Hið Íslenzka Fornritafélag, 1933), pp. liii–xcv. While this condition is common among medieval texts, the fact that it should be true for an entire genre raises a number of pertinent questions and possibilities. The names of many skaldic poets are diligently recorded in the sagas, so it might be that saga authorship was understood to be either a lower artistic form or one less indebted to the creative agency of an individual when compared with skaldic verse. The scribes who wrote or copied the written sagas may have thought of themselves (or wished to present themselves) as recorders of an extant oral tradition rather than innovators of original written works. A further possibility is that sagas were not considered to be literary works, but something more akin to chronicles, and therefore authorship was thought to matter less. The reasons for this ubiquitous anonymity are frequently debated.4444Andersson, Growth of the Medieval Icelandic Sagas; Clover, Medieval Saga; Gabriel Turville-Petre, Origins of Icelandic Literature (Oxford: Clarendon Press, 1953). For our purposes, however, the fact of this genre-wide anonymity is telling in itself. By removing the intermediary figure of the author from the frame, the sagas give the impression of speaking with a common, corporate voice. This sense is reinforced by their frequent habit of intertextually referencing one another — for instance, in Laxdœla saga: ‘Gunnarr hafði sekr orðit um víg Þiðranda Geitissonar ór Krossavík, sem segir í sǫgu Njarðvíkinga’ (Gunnarr had been outlawed for slaying Þiðrandi, Geitir’s son, of Krossavík, as is told in the Saga of the People of Njarðvík).4545Laxdæla saga, ed. by Einar Ólafur Sveinsson, Íslenzk Fornrit, 5 (Reykjavik: Hið Íslenzka Fornritafélag, 1934), p. 202. Or, in another instance, in Þorskfirðinga saga: ‘Þeir Guðmundr félagar urðu sárir nökkut, ok fóru þeir utan um sumarit, sem ætlat var, ok er mikil saga af þeim í Nóregi frá viðskiptum þeira Ölvis hnúfu’ (Guðmundr and his companions were somewhat wounded, and they travelled to Norway that summer, as they had intended, and there is a great saga about them in Norway and their dealings with Ǫlvir Hump).4646Þorskfirðinga saga, in Harðar saga: Bárðar saga, Þorskfirðinga saga, Flóamanna saga, ed. by Þórhallur Vilmundarson and Bjarni Vilhjálmsson, Íslenzk Fornrit, 13 (Reykjavik: Hið Íslenzka Fornritafélag, 1991), pp. 173–227 (p. 226). Or, in a third case, in Grettis saga: ‘þaðan af gerðisk saga Bǫðmóðs ok Grímólfs ok Gerpis’ (the Saga of Bǫðmóðr, Grímólfr, and Gerpir describes the events that followed).4747Grettis saga Ásmundarsonar, p. 32. No saga of this name exists in the surviving corpus, so we may assume that it was either lost or never written. What emerges from these examples is the sense of a network of literate authors attempting to create the impression of a unified saga corpus which, in order to be fully understood, needs to be accessed as a whole. The naming conventions used in the first and last of these examples suggest the existence of discrete sagas, whether written or oral, known by those names and sufficiently static that each author could be confident the incident they were alluding to would appear in every redaction to which their audience might have access. This device reinforces the imperative for readers to accept the emerging authorized canon of written sagas, since divergent oral iterations might not supply these cross-references so reliably.

As the Middle Ages wore on and the written word attained primacy over the spoken word in Iceland, a process of closure took place by way of which an authoritative canon of saga variants began to enter circulation. Part of this process involved the appending of prolegomena to written sagas. These used a range of rhetorical strategies to stress the authenticity of the particular redactions which they introduced and concluded. Where oral discourse was open and fluid, the very practice of containing and constraining the sagas — ‘that which is said’ — on the page involved the generating of a hierarchy which sought to privilege the written saga over any competing spoken traditions; the closed book over the open oral exchange. Despite the strong imperatives in favour of the written saga as a means of imposing editorial control on the form, this process of textualization and its effects were gradual and piecemeal. While the act of inscribing a particular redaction on parchment closed down certain possibilities for the simultaneous circulation of several mutually contradictory yet equally authoritative versions of a given saga, the existence of saga manuscripts written in the vernacular also enabled a process of opening up the corpus, which had previously been enclosed in the minds and memories of a knowledgeable few, to successive generations. Since the primary mode of reception remained oral well into the high Middle Ages, through a process of listening as sagas were read aloud, the writing down of sagas opened a new range of performative possibilities couched in the interplay between the voices of the author, scribe, compiler, reader, and, potentially, the contributing voices of audience members. Far from fixing the sagas in a static form, these manuscripts might be better understood as vessels containing the fluid stories for a time, but ultimately intended to be opened up with each rereading to a new generation of Icelanders.