Manuscript Hunting in Late Seventeenth-Century Iceland

The fervent collecting of ancient and medieval manuscripts—in Italy and Greece from the late fourteenth century, in France, Germany and England from the early fifteenth century, and in Spain and Iceland from the late sixteenth century onwards—resulted not only in the accumulation of new texts and information. It was also a reason for perplexity as scholars struggled to design methods for sifting evidence, for defining options, and for making choices concerning the texts they wished to use or publish, their struggle at times ending in despair and exhaustion. As more and more manuscripts were brought to light and made accessible in continuously growing libraries and private collections, scholars needed new tools to understand and manage textual variance. Which manuscripts should be used, and how could their quality be gauged? Why did their texts differ? How should they be transcribed and published? Should vernacular texts be translated into Latin? What kind of comments were needed – factual, historical, or textual?

Such technicalities, in other words the painstaking travails of textual scholarship, deserve to be a focus of attention in an overarching history of the humanities. This history should not restrict itself to the brilliant ideas of philosophers and theologians. Developments in the humanities were based on the “rough” material of ancient textual sources, and in this article I hope to show how scholarship in the Early Modern age needed to plough through vexed issues of textual variance, manuscript hunting and compilation that, although often coming to dead ends and not always resulting in successful publications, were nevertheless seminal to developments in history writing, philology, and language scholarship.

I will focus on an area to which the history of scholarship has paid scant attention: the kingdom of Denmark-Norway in the late seventeenth century. Somewhat earlier, the English antiquarian Henry Spelman characterized this “unlearned” kingdom as being on “the confines of the Arctic Circle.”

Although an abundance of Icelandic and Norwegian medieval texts and manuscripts was rediscovered and the first editions appeared in 1664-1673, none of them was based on a critical assessment of manuscripts. In the 1680s and 1690s a renewed effort was made that, had it succeeded, would have delivered interesting results. Lamentably, however, this endeavour did not produce editions, partly due to a lack of funds but chiefly because of methodological difficulties. The scholars involved were aware of the value of vellum manuscripts but unable to solve the problem of textual variance, that is the fact that manuscripts of any specific work differed in many ways, and a scholar who wanted to produce an edition would have to make choices based on critical discernment and clear principles.

Instead of using standards of philological correctness of our present age to judge these late seventeenth-century efforts, I shall base my estimate on the exceptional insights of the late fifteenth-century Florentine scholar Angelo Politian, the enfant terrible of late medieval and early modern scholarship. In a way, Politian’s philological ideals were so close to modern procedures that they seem almost too good to be true; his practice in that sense resembles what Ezio Ornato has calleduna semplice curiosità archeologica” [a mere archaeological curiosity]. Politian’s example was only followed by scholars in more recent times, but nonetheless his philological ideal can be used as a benchmark for developments in the Early Modern period. Arrogant and impatient at times, he was an incredibly observant scholar who wished to outperform his contemporaries as he strove to establish better versions of ancient Greek and Roman texts through linguistic refinement and the meticulous observation of manuscripts. He wanted to base his work on a thorough investigation of as many manuscripts as possible, preferring older ones but not discarding more recent texts that might be copies of ancient books.

Politian’s work demonstrates his critical assessment of the quality of manuscripts and the ways in which they were related. He made very careful collations of manuscripts and printed editions, rigorous transcripts of texts that he borrowed, and emendations, judiciously based on other texts written by the same author or his direct environment. Politian loathed inexactitude and harshly criticized his predecessors and contemporaries for making too many mistakes. In his Miscellaneorum centuria prima of 1489, for one, he attacked the late Domizio Calderini for his sloppy method. Jacopo Antiquari, Calderini’s friend in Milan, complained about this in a letter and claimed that since Calderini was long dead, it was like attacking a ghost. Politian retorted on 30 November 1489 that he could not know whether Calderini would have corrected his mistakes if he had still been alive:

How should I know this? Are you really saying that it is always and everywhere a good idea to expect intellectual progress? [An usquequaque de ingeniorum profectu bene sperare est?] Were not all of this particular person’s final contributions even more inaccurate?

In Politian’s view, progress came with hard work. His influence, however, was limited—like Calderini, in this respect, he died too soon. Most scholars continued to make many mistakes, and even the best were careless in their choice of manuscripts and variants. When Erasmus, for instance, published his Greek edition of the New Testament in Basel in 1516, he used five manuscripts. One of them is preserved, a twelfth-century vellum now known to be devoid of textual value. Although aware of the value of old manuscripts, Erasmus relied mostly on recent ones with scant merit. Where the Greek text was lacking, he even added bits of his own translation from the Vulgate, thus revealing, to quote L.D. Reynolds and N.G. Wilson, “a lack of a set of logical principles for the evaluation of manuscripts.” In his edition of the works of Seneca in 1515 and 1529, Erasmus recognized the worth of a ninth-century manuscript he had acquired, but he did not use it much: “Instead of basing his text of these works upon this prime witness, he drew on it spasmodically to emend what he had before him.”

It soon became standard practice for most editors to use first editions as a basic text—textus receptus or textus vulgatus—adding selected readings from manuscripts that were mentioned only vaguely (“emendatio ope codicum”) and various conjectures (“emendatio ope ingenii”), at times original, but just as often borrowed or stolen. Th ere were some exceptions, however. In around 1530-1570, scholars of Roman law, such as Antonio Agustín, Piero Vettori, Jacques Cujas and Joseph Scaliger, criticized older editions severely, tracing the genealogy of manuscripts and publishing transcripts of those they considered important. They normalized orthography and corrected errors, and even made typographical distinctions between variants and conjectures.

Such efforts were consolidated every time that scholars made a determined effort to gather manuscripts in growing libraries. Although most libraries were private, the biggest ones, kept by kings, dukes and universities, were public. From the 1640s onwards, this enabled Dutch scholars such as Nicolaas Heinsius and Isaac Vossius to produce better editions, partly through their ingenuity in making emendations and conjectures based on a wider knowledge of texts, but also through a closer scrutiny of manuscripts. Scholars came to prefer their specific texts above earlier editions, although at times some were overwhelmed and only produced heaps of variants without distinction.

This process of strenuous and uneven progress was mirrored further to the north, but developed more slowly and with less spectacular results in terms of editions. Influenced by Italian humanists, scholars in Northern Europe had begun searching for documents and manuscripts already in the early sixteenth century. Interesting texts were discovered such as the plays and poetry composed by Hrotsvitha von Gandersheim in the tenth century; they were edited and published by Konrad Celtis in Nürnberg 1501.

In 1515, his friend Konrad Peutinger used recently found manuscripts to publish The History of the Goths, written by Jordanes in the sixth century, and The History of the Longobards by the ninth-century historian Paulus Diaconus.

In 1514, Christiern Pedersen, a Danish student of theology in Paris, published the extensive early thirteenth-century Latin History of the Danes (Gesta Danorum) by Saxo Grammaticus—and was actually among the first editors to use the blurb text for claiming that this was the first time a text appeared in print: “nunc primum literaria seriæ illustratæ tersissime que impressæ.” Later, he wished to make a Danish translation of Saxo with detailed explanations on Nordic history. Aware of the existence of manuscripts of medieval sagas of Norwegian kings, he hired a Norwegian man of learning who knew the old language to make excerpts.

A Danish translation of Saxo first appeared in 1575, when the historian Anders Sørensen Vedel concluded one. The first Danish translation of the Norwegian kings’ sagas appeared in Copenhagen in 1594 and a more thorough one in 1633, when Ole Worm, in his introduction, compared the writings of the thirteenth-century Icelandic historian Snorri Sturluson to Greek and Roman historical works, stating that they were just as useful and truthful.

As late as the 1630s, the Dutch historians Johannes Meursius and Johannes Pontanus wrote eloquently on the medieval history of Denmark without using a single manuscript. It was by then well known that medieval manuscripts of great interest for the medieval history and culture of Scandinavia were to be found in Iceland. The first batches were sent as gifts to the Danish king in 1656 and 1662 by the learned bishop Brynjólfur Sveinsson who hoped that Icelandic texts would be published with a Latin translation and a scholarly commentary.

Swedish scholars, arguing with Danish ones about the age of their two nations, also showed an interest in Icelandic texts containing references to Swedish kings. In 1664-1673, a series of text editions appeared in Uppsala and Copenhagen with translations into Danish, Swedish and Latin, but all of them were worse than anything produced in Europe at the time. Any available manuscripts were used regardless of their scholarly value, mostly recent and inexact copies from Iceland, and the editors’ commentaries reveal a credulous attitude towards the text and a blatant lack of critical zeal.

This was the scholarly norm when Árni Magnússon [1663-1730], or Arnas Magnæus as he will be called here, arrived in Copenhagen in 1683, at the age of twenty, to study theology at the university, just like many promising young men of good families in Iceland. A year later he became assistant to Thomas Bartholin the Younger, the recently appointed royal antiquarian, and provided numerous Icelandic texts for his fairly voluminous book on the fearlessness of medieval Danes, published in Copenhagen in 1689.

Magnæus developed rapidly as a textual scholar. In 1684-1685 he made hundreds of short excerpts from Icelandic manuscripts of sagas and historical works, using recent copies available in the city, most of them of disputable quality, copied without care for reading purposes and not for scholarly use. The Icelandic vellums that belonged to the king and the university were all at Stangeland, on the west coast of Norway, with the Icelander Þormóður Torfason or Thormod Torfæus, royal historiographer of Norway.

Magnæus’s method of transcription was simple. He used his own orthography and expanded the abbreviations in the manuscript he copied (exemplar), so that he gave everything his personal touch. Bartholin and Magnæus did not care about the age of the manuscripts or texts they used. To them, all texts were equally interesting, and they used what was at hand, such as an early seventeenth-century copy of Knytlinga saga, a saga on Danish medieval kings that Ole Worm had received from Iceland some decades earlier. In collaboration with an unknown Icelander, Magnæus copied the legendary Hrólfs saga kraka from a paper copy and translated it into Latin. He also made a copy of the Norwegian thirteenth-century Speculum regale (Konungs skuggsjá), again in collaboration with an unknown scribe, but now from a vellum manuscript that belonged to Peder Resen, professor of law at the university. In one place, Magnæus commented on the fact that a leaf was missing. He corrected what the other scribe had done, and they both modernized the spelling.

Although one of his exemplars was on vellum and the other a recent copy, the method was consistent. The goal was to make texts readable and accessible, so that information would be available on a variety of relevant issues. The transcripts were meticulous, and few changes were made, except for the spelling. This is how Icelandic scribes had worked earlier in the century. Magnæus followed their tradition and not without success.

In 1685-1686, Magnæus changed his mind completely, as if he suddenly discovered that things could be done so much better. In the spring of 1685, he went to Iceland in order to collect manuscripts for Bartholin; he stayed until the autumn of 1686. He found nothing that Bartholin needed, but obtained for himself a few legal codices and made copies of medieval legal texts in a manner totally different from what he had done so far, in more detail and with greater exactitude. As he returned to Copenhagen, he responded to this “revelation” by transcribing several texts from a late fourteenth-century collection of Icelandic sagas, recently acquired by Resen from a student who had been with Torfæus as a scribe for three years.

The transcript is exceptionally detailed and can be situated on what philologists now call the diplomatic level. Magnæus strove to imitate the orthography and abbreviations of the original, and succeeded except for minor inconsistencies in some details. He emended the text in a few places and put his additions within square brackets. This “new” method seems to have been based on a mixture of foreign models and personal fascination due to direct contact with vellum manuscripts.

The exceptional attention to detail was probably inspired by his reading of scholarly books and editions in Bartholin’s library of at least 2400 volumes. Here, Magnæus had access to hundreds of editions of classical and medieval texts and to the most recent scholarly feats of Jean Mabillon and others—not to mention Angeli Politiani opera, published in Paris in 1519. By now, Magnæus had decided to acquire as many Icelandic texts as possible in good copies, probably hoping to establish a research library of sorts. However, he soon found the newly developed diplomatic method too time-consuming and opted instead for a somewhat easier way of copying texts in a normalized orthography with expanded abbreviations. He did not go so far, however, as to use his own orthography like he had done earlier. In the winter of 1687-1688 he copied another fourteenth-century vellum of Icelandic sagas, Möðruvallabók, with two friends who studied at the university, Ásgeir Jónsson and Eyjólfur Björnsson. They expanded all abbreviations and wrote all names with a majuscule. Corrections were not indicated, and not all medieval letter-forms were maintained.

In the autumn of 1688, Torfæus came to Copenhagen to claim his salary that was long overdue. He was working on books on the medieval history of Denmark and Norway, published in 1702 and 1711, respectively. He befriended his much younger countryman Magnæus who confided that he planned an edition of a crucial historical work, the early twelfth-century Book of Icelanders (Íslendingabók). Written by chieftain, priest and historian Ari Thorgilsson the Wise, it succinctly narrates the discovery and colonization of Iceland in the late ninth century and its subsequent Christianization. The edition was intended for scholarly readers and was to contain a Latin translation and an extensive historical commentary. It would have added invaluably to the knowledge of Icelandic medieval history and was accepted for publication in Copenhagen in spring 1691, but never appeared. My contention is that Magnæus abandoned the edition because textual and factual inconsistencies in medieval manuscripts stretched the limits of his method. Rather than accepting these inconsistencies as some sort of challenge by integrating them in his work, he discontinued the project.

Magnæus’s first problem was that he had no decent copy of the text and had to use whatever manuscript he could lay his hands on. To our present knowledge, an early thirteenth-century vellum manuscript was extant in Iceland in around 1650. A highly qualified scribe, Jón Erlendsson, made two copies for the aforementioned Bishop Brynjólfur before the manuscript disappeared, one of them (AM 113 b fol.) better than the other (AM 113 a fol.)—now both in the Arnamagnæan collection. Several copies were soon made of the version that was most faulty; subsequent copies of these copies grew more deficient with each generation. At some point in 1688, Magnæus received one of those bad copies, unfortunately not preserved, and made a copy for himself that he intended to use in his edition.

In the summer of 1688, the Book of Icelanders was published together with the much longer Book of Settlement by Bishop Thordur Thorlaksson at the only printing press in Iceland. The text was based directly on the manuscript AM 113 a fol. with a few judicious corrections. The editorial principles are explained in a short introduction, but there is no commentary and no translation. When Magnæus received the printed book in the autumn of that year, he must have realized that this edition was far better than the copy he had made for himself. Paradoxically, he showed no sign of using the edition in his work but instead insisted that Torfæus should send him a manuscript of the text even if it contained exactly the same text as the printed edition.

In early 1690, Magnæus added to his copy a few marginal notes based on this manuscript, but should have made many more had he wanted to be thorough and consistent. In 1691 and 1692, Magnæus received both of Jón Erlendsson’s mid-century copies from Iceland and corrected his text in various places on the basis of the better copy, but again not consistently. Simultaneously, he collated that copy with the other good one. He now had the best manuscripts and knew it, but as he saw that there were more and different copies, he seems to have lost his way and instead of redoing the text, he abandoned the project. Magnæus’s other problem was the commentary. The Book of Icelanders is only five thousand words in length, but poses intricate issues of chronology that appeared insoluble—and some of them really are. Magnæus went to stay with Torfæus at Stangeland for a few months in 1689.

On his way back to Copenhagen, he wrote a letter asking Torfæus where he had read that the Norwegian king Harald Fairhair was born at the moment that King Gorm of Denmark had been in power for seven years. Torfæus replied that this conclusion was “ex hypothesi,” since medieval chronicles claimed that Gorm became king in the year 840 and Harald was born in 848. The problem Magnæus and Torfæus encountered was that medieval chronicles and kings’ sagas hardly ever mention dates and do not agree with each other on how many years various kings stayed in power.

About Harald Fairhair, allegedly the first king of a unified Norway, Magnæus concluded at one point that there was so much confusion that he saw no possibility of figuring out anything at all. Torfæus claimed that only the oldest historians should be used and not the ones that followed them. No author, however, could be entirely trusted since all of them made their own guesses and even mixed things up. For this reason, Torfæus repeated, a scholar had to work “ex hypothesibus.” Having done what could be done, the scholar should decide, or better still the community of scholars, or as Torfæus wrote on 2 October 1690: “It is best that both of us agree, and Bartholinus, and exclaim a single adieu!”

When Magnæus insisted, Torfæus offered to include his arguments in the book. Magnæus should also publish both versions, and it would be left to the readers to decide. Any further discussion would be a waste of time and paper. Torfæus went on to write several learned volumes, discussing various options for a variety of problems in the texts, but Magnæus only produced handwritten notes. It needs to be said that a contributing factor to Magnæus’s failure to produce an edition of the Book of Icelanders may have been the perceived or real lack of interest in Nordic medieval history on the continent. This lack of interest is apparent, for instance, in the fact that the only writers of Danish history mentioned in Adam Rechenberg’s general presentation of necessary learning, published in Leipzig in 1691, were Saxo Grammaticus and Pontanus, whose Rerum Danicarum historia was published sixty years earlier.

Perhaps in response to this lack of interest, Magnæus published a short Danish medieval chronicle in Latin in Leipzig in 1695, transcribed from an old vellum manuscript (“pervetusto codice membraneo“) that belonged to the university library in Copenhagen. In his introduction Magnæus shows his debt to his predecessors by connecting his modest effort to famous names in the scholarly world, praising Marcus Meibomius, Roger Twysden and Jean Mabillon for their exemplary editions. His hope of publishing a manuscript fragment on Danish kings did not come true, however. In his own words, this was because of the reluctance of German printers to print a text in the Icelandic language, although it is clear that his Latin translation was far from ready and the commentary in shambles, as Magnæus constantly changed his mind, and he was, as with The Book of Icelanders before, overcome with doubts.

The lesson Magnæus learned from his erudite struggles was that instead of writing and publishing, he should do two things, both of them within what can be called Politian’s program of a necessary assessment of the quality of manuscripts and the making of rigorous transcripts of borrowed ones:

1. Transcripts should be made with great care and exactitude. One example of this notion will have to suffice, relative to two Icelandic vellum manuscripts in the Ducal Library in Wolfenbüttel, registered as Icelandic mythology and poems. In the early months of 1697 the dukes sent their librarian [Lorenz] Hertel to Stockholm in order to pay their respects to Sweden’s new king. Gottfried Wilhelm Leibniz, philosopher, historian and head-librarian in Wolfenbüttel, had recently become interested in septentrional matters and sent one of the manuscripts to his correspondent, the Swedish scholar Johan Gabriel Sparwenfeld. Leibniz did not know the contents but thought that the manuscript was written in verse without rhyme “en Islandois ou en vieux Scandinave.” Sparwenfeld persuaded Hertel to take the manuscript to Magnæus in Copenhagen, something Leibniz agreed to afterwards:

Mr. Hertel left our Icelandic manuscript with Mister Arnas Magnæus in Copenhagen, if it is useful to him, so much the better. Libraries and manuscripts need only be available for the use of skilled people.

Magnæus had the aforementioned Ásgeir Jónsson, again in Copenhagen, make a copy of the two Icelandic thirteenth-century family sagas: Eyrbyggja saga (AM 450 a 4to) and Egils saga (AM 461 4to). As the first and last pages were illegible, Magnæus copied them himself. He already had several copies of both sagas and filled two gaps in the Wolfenbüttel manuscript with a text from other versions. Another scribe helped him to compare the transcript with the original, and Magnæus concluded that, although not copied letter by letter and except for the poems and some important sentences, the transcript was reliable.

In June 1701 he sent the manuscript back with a report on its contents. In his collection as it survives to this day, there are at least 500 copies that he made or had his assistants make, all of them based on these principles, besides thousands of transcripts of Icelandic, Norwegian and Danish charters and other documents.

2. All extant manuscripts should be tracked down. Many interesting and quite dramatic stories could be told about Magnæus’s search for manuscripts. A good example is his chase after a fourteenth-century manuscript of the historical work Sturlunga saga, only preserved there and in another contemporary manuscript, acquired by Magnæus in 1699 (AM 122 a fol.).

The so-called Reykjarfjarðarbók (AM 122 b fol.) was torn to pieces in 1676-1679 as it was damaged by moisture. The owner, a well-to-do farmer, gave leaves to his friends for use as book covers. Magnæus was informed about the manuscript’s tormented existence in 1693 and went on to trace its remains for three decades. In all, he retrieved 30 pages out of an estimated 180 that made up the original manuscript. At least seven colleagues sent him one or more leaves, the last two leaves arriving in 1724. His painstaking efforts can be seen in his request to Árni Gudmundsson in 1707, as he asked where and when the first known owner at Reykjarfjörður had obtained the book. Was it complete when Árni saw it? If not, what was missing: how much or how little, in the beginning, middle or end? Was the whole book readable or was it damaged because of moisture or black stains? Who exactly possessed leaves, and would it be possible to retrieve them?

When Magnæus died on 7 January 1730, his collection comprised close to three thousand manuscripts and fragments, almost one-third of them on vellum. Half of those are fragments of less than six leaves. In a limited sense it can be claimed that he thus came close to the idea of recension—fully developed in the late eighteenth century, according to Timpanaro—as he desired to gather together all manuscripts of a determined text, not only vellums but also seventeenth-century copies. In this wish for completeness, Magnæus went further than any contemporary scholar or collector of manuscripts. Unlike Politian, however, he never developed a clear sense of how to manage the difference between manuscripts. He was not alone in this, as shown from the careless and arbitrary method of Lenglet de Fresnoy in his edition of Le Roman de la Rose in 1735; but it must be said that some or even many editors of that time were more scrupulous.

In the course of his studies, Magnæus’s understanding of textual variance hardly improved, as can be seen in his handling of some seventeenth-century copies of an otherwise lost medieval text on church history called Hungurvaka. It was hopelessly confused, and I shall spare the reader the details. Before 1700, one of his assistants made a copy (a) of one (b) of these manuscripts and Magnæus collated it with another one (c). He then asked another assistant to make a copy (d) of his collated copy (b), but without the added variants (“varias lectiones”). In 1724, Magnæus compared this second copy (d) “accuratissime” with the original (b) used to produce the first copy (a), making numerous corrections. Finally, he collated his copy (d) with a third seventeenth-century manuscript (e) and wrote down all the differences in the margins. His copy now contained an “accuratissima collation” of the two oldest manuscripts (b, e) of this important text, as he happily concluded. The only reason for doing all this appears to have been to throw away the first copy (a), useless since he now had its text on the margins of other copies.

Returning to the Book of Icelanders in his later years, Magnæus concluded, erroneously as we have seen, that as a young man he had copied the text as exactly as he then could. In around 1720 he made an exact copy of his best copy and most likely recognized that his own version of 1688-1690 was useless, although he did not discard it. He copied some of his old commentary and revised a number of items. When, ultimately, he did not publish an edition, he now justified his inactivity by claiming that the world of books was replete with products of vanity. He had never intended to write books himself and was convinced, as he explained to an assistant, “that a man could spend almost his whole life in putting together a little booklet.”

As a philologist, Magnæus gradually gained a finer understanding of the quality of texts and manuscripts, but this understanding was never brought to fruition in printed editions. His scholarship, seminal as it may have been for the origin of Icelandic studies, did not reach the public arena of European philology. The Republic of Letters allowed the participation of Scandinavian scholars of course, but there was limited interest in northern languages, culture and history. The thriving Anglo-Saxon studies in England were the closest field of research, much of it just as admirable in the details and the driving force the same, that is to replace or at least displace the old view that the origin of European languages and culture should be sought in Greek or Roman Antiquity. There was little contact, however, and scholars of medieval Iceland remained isolated.

We can thus safely conclude that Magnæus’s diligent but inconspicuous scholarship, inspired as it may have been by the wish to transplant the ideas of humanistic philology to Scandinavia, was only partly successful in the sense that his activities did not become an integrated part of the wider European developments in philology and historiography. His image in the historiography of the humanities therefore remains that of a hunter for manuscripts and maker of copies, and he left the task of coping with textual variance to posterity by donating his collection to the University of Copenhagen after his death. To be fair, some of the problems he encountered remain unresolved to this day.


Már Jónsson is professor of history at the University of Iceland (Reykjavík). This article originally appeared as a chapter in The Making of the Humanities. Volume 1: Early Modern Europe.


Featured: etching of Árni Magnússon (1663-1730), Icelandic scholar and collector of manuscripts.