SVALBARÐS FUNDR. THE PLACE NAME SVALBARD AND ITS CONNOTATIONS IN MEDIEVAL AND MODERN LITERATURE AND CARTOGRAPHY1

The paper analyses medieval forms of the name Svalbard as applied to the land “discovered” in 1194, suggests that this Arctic discovery could have been named after a farmstead in Iceland, and follows the story of the name by discussing its contexts in medieval and modern literature and on maps. However little information about Svalbard survived in the Icelandic annals, the Landnámabók, and related texts, it became part of competing visions of the Arctic, from the late medieval Samsons saga fagra through the adoption of Svalbard as the name of a new territory under Norwegian rule in 1925.

The official change of the name of a group of islands in the Arctic to Svalbard in 1925 stressed the continuity between late twelfth-century seafarers, the assumed first discoverers of the archipelago, and modern Norwegian explorers, whose names had been added to the map of Spitsbergen in the process of its Norwegianization. The bulk of research on the medieval name Svalbard and its meaning was done during heated debates of the early twentieth century, when Norwegian scholars became political actors actively involved in claiming Norway's historical rights over Spitsbergen.2 Today, when the issue of sovereignty has been long settled, a general skepticism towards the philological and historical conclusions of these scholars has become the norm in serious research both inside and outside of Norway. The standard Norwegian textbook on the history of (Norwegian) Svalbard devotes only two pages to its medieval eponym and concludes that the identity of the Svalbard discovered in 1194 and the group of islands renamed Svalbard in 1925 is an issue of "faith and doubt" (Arlov 2008, 50-51). A foremost German authority on Scandinavian geography mentions the Arctic island of Svalbarði only in passing; in one publication he identifies it as "Jan Mayen rather than Spitsbergen" (Simek 1986, 252, 258), and, in another, unambiguously equates it with Spitsbergen (Simek 1990, 187).
Aside from the general skepticism about or tacit acceptance of the identity of medieval and modern Svalbard, there has been little discussion of the medieval name and its context since the mid-twentieth century. Few publications have even sought to list all contexts in 1 I am grateful to Elena Aleksandrovna Melnikova and to Nordlit's anonymous peer reviewer for their comments on the first draft of this article. 2 The history of political and toponymic Norwegianization of the Arctic has been outlined in Berg 2013; on politically motivated scholarship in the period leading up to Norwegian sovereignty, see Drivenes 2012, who also noted (55) that mapping activities of the expeditions of 1906 and 1907 resulted in a heavier presence of Norwegian names in Spitsbergen's toponymy; on the modern toponymy of Svalbard, which includes the name of Spitsbergen as its largest island, see: Place Names of Svalbard 2003. which the medieval name appeared (Krawczyk 1987;Starkov 1998), let alone discuss the name's variants and forms. "Svalbard" does not have an entry in any of the principal modern international reference works devoted to the European or Scandinavian Middle Ages, such as Lexikon des Mittelalters, Medieval Scandinavia, or the last printed edition of the Reallexikon der Germanischen Altertumskunde (B. 30, 2005).
The main purpose of this article is to take stock of the forms and contexts in which the name "Svalbard" appears in narrative texts and cartography. The survey begins with reports on Svalbard in the Icelandic annals, which provide us with the date of its discovery (the article's first three sections), and in the Landnámabók, which gives directions and distances to Svalbard from the Langanes peninsula in northeast Iceland (the fourth section). This article also discusses various identifications of Svalbard in relation to Greenland and Northern Eurasia provided in Samsons saga fagra (the fifth section) and in the geography and cartography of the seventeenth, eighteenth, and nineteenth centuries (the last three sections). The previously neglected material introduced in the first half of this article may help identify the elusive land discovered in 1194. Other contexts, especially the scholarly works and maps analyzed further on, are not useful for uncovering the identity of the "real" Svalbard, but they are interesting as attempts to match ancient Scandinavian geographical lore with contemporary discoveries. They helped to keep the name "Svalbard" afloat, sometimes in unexpected areas of the Arctic.

*Svalbarð and its Icelandic namesakes
The nominative form *Svalbarð from which the modern Norwegian name of the archipelago is derived, does not occur in medieval texts related to the discovery of 1194. It is a modern reconstruction based on the genitive form Svalbarðs, which can, however, be used to reconstruct not only a neuter name Svalbarð, but also a masculine name, Svalbarðr. The genitive form Svalbarðs appears in different orthographic variants in several versions of the Icelandic annals in the terse entry for 1194: Svalbarz fundr, Svalbarðs fundr, Svalbarðzfvnndr, Sualbards fvndr (Islandske annaler 1888, 22, 62, 121=477, 324), which literally means "Svalbard's discovery", that is, the discovery of Svalbard. The earliest manuscript with this form (and the earliest manuscript with this name overall) is Annales regii (Konungsannáll) mostly written about 1300 (Stofnun Árna Magnússonar í íslenskum fraeðum, GKS 2087 4to).
One argument in favor of the name's neuter gender is its generally accepted etymology. The first component sval-refers to the adjective with the meaning "chilly, cool", while one of the meanings of the second component, the word barð, is "edge, rim". As the nominative case of the neuter Norse noun is barð, it is quite logical to reconstruct the nominative case of the toponym as Svalbarð.
The second argument is the existence of similar toponyms in Iceland. Neither the distribution nor chronology of this toponymy has been studied, although such a study would obviously impact our understanding of the 1194 discovery if it could suggest that this discovery was named after an Icelandic eponym. It would increase the probability that the discoverers were Icelanders rather than Norwegians (on this issue compare Finnur Jónsson 1926, 81).
Svalbarð is a name of several farmsteads in different parts of Iceland, namely in Dalasýsla; on the western side of Vatnsnes peninsula in Vestur-Húnavatnssýsla; in Norður-Þingeyjarsýsla; and in Suður-Þingeyjarsýsla. One of the farmsteads is close to Langanes peninsula in Norður-Þingeyjarsýsla in northeast Iceland, from where, according Svalbarðs fundr. The Place Name Svalbard and Its Connotations in Medieval and Modern Literature and Cartography 20 to the Landnámabók, the trip to the Arctic Svalbard began. Located on the other side of Þistilfjörður bay, on its southwestern coast (at 66°12'26"N 15°43'04"W), this Svalbarð is an ancient settlement. Its earliest appearance in written sources is in the book of property records of the diocese of Hólar collected by bishop Auðunn in 1318 (see Diplomatarium Islandicum 1893, 2: 425-426, Eiríkur Þormóðsson 1972. The archaeological evidence indicates, however, that the farm or at least the farm midden existed earlier (Amorosi 1992, 104), from the early eleventh century. Another Svalbarð, found on the eastern coast of Eyjafjörður (at 65°44'40"N 18°05'02"W) in the neighboring county of Suður-Þingeyjarsýsla, is an even older and far more famous farmstead (see: Saedís Gunnarsdóttir 2005, 61-70). The Landnámabók indicates that it had been founded sixteen years before the Christianization of Iceland in 884 (Íslendingabók. Landnámabók 1968, 2: 269-270 Jónsson 1908, 181;Lind 1920-21, 370). The church of Svalbarð was described in the same 1318 book of property records (Diplomatarium Islandicum 1893, 2: 448-449). Both Svalbarðs are coastal areas with hills in the background, which, though not unusual topography for an Icelandic farmstead, could support the interpretation of the name Svalbarð's second component as "edge, rim". As the meaning of the first component ("chilly, cool") is concerned, we can see from the two Svalbarðs' vibrant history that the cool climate did not render them uninhabitable.3 Because of their proximity to Langanes and relative importance in the larger region, the newly discovered land could have been named after one of the farmsteads and its topography could have resembled that of the Icelandic Svalbarðs.

Svalbarðr in the Guðmundar saga
If we set aside the etymology and ignore the gender of both the Icelandic word barð and the Icelandic farm names, we can use the genitive Svalbarðs to reconstruct a masculine nominative singular form Svalbarðr. Indeed, this was how the name was reconstructed for the index in Gustav Storm's edition of the Icelandic annals (Islandske annaler 1888, 637).
The only nominative singular form which is attested by medieval sources and corresponds to the genitive singular Svalbarðs, is in fact a form with the masculine morphological marker -r. This masculine form Sualbarðr appeared in recension A of the Guðmundar saga, an account of the life of Guðmundr Arason (1161-1237), Bishop of Hólar from 1203 to 1237. Stefán Karlsson dated the main manuscript of this recension (Københavns Universitet, Den Arnamagnaeanske samling, AM 399 4to) between 1330 and 1350 "in all likelihood" (Guðmundar sögur biskups I 1983, XLI).
One of the principal sources of recension A of the Guðmundar saga was a version of the Icelandic annals from which Bishop Guðmundr's biographer borrowed the fragment mentioning the discovery of Svalbard together with the news of King Sverrir's coronation (29 June 1194), which precedes the fragment in most versions of the annals: Vigðr Sverrir 21 konungr undir korono. Þa fannz ok Sualbarðr (Guðmundar sögur biskups I 1983, 83). As compared to most versions of the annals, the noun fundr, "discovery" was replaced by the verb fannz (fannst), the past of finnast, a mediopassive derivative of finna, "turn up" in the sense of "be found". Accordingly, the meaning of the new phrase is "Also, Sualbarðr was found at that time".
Why should this name, contrary to its assumed etymology, have received the nominative ending -r which marked masculine in Old Norse? One explanation for the form Sualbarðr in recension A of the Guðmundar saga is that the place name may have been reanalyzed as a personal name analogous to Hárbarðr, Rádbarðr or Langbarðr (where the second component almost always refers to barð, "beard"; see Sturtevant 1952, 1152-1153, Hødnebø 1972. Interestingly, this was the time of a wholesale change in the assignment of grammatical gender in the continental Scandinavian languages, when the default gender shifted from neuter to masculine (see : Steinmetz 2006;Trosterud 2006). This shift followed, however. the loss of the masculine nominative ending -r. Thus, the word bard "edge, rim, (river) shore" is masculine in New Norwegian, which may explain why Gustav Storm did not see a problem with masculinizing the Old Norse name as Svalbarðr.
The nominative Svalbarðr occasionally appears in literature, not necessarily following recension A of the Guðmundar saga or the index to Storm's edition of the Icelandic Annals. Thus, E.V. Gordon, who, in the preface to his authoritative textbook, specifically stressed the careful attention he paid to the morphological marker -r (1927, VIII), rendered the name of "the last northern discovery of the Norsemen" as "Svalbarðr, 'the cold edge'" (XXVII-XVIII). Despite criticism by a reviewer who could not understand "where the author got this form" (Stefán Einarsson 1928, 543), Svalbarðr can still be found in modern reference tools (Murdoch and Read 2004, 13).

Svalbarði in Skálholts annáll
Another variant, Svalbarði, appears in yet another version of the same 1194 entry in one of the Annals manuscripts, Skálholts annáll of ca. 1362, which rephrases the statement Svalbarðs fundr with Svalbarði fvndinn (Stofnun Árna Magnússonar í íslenskum fraeðum, AM 420 a 4to, fol. 7r;4 Islandske annaler 1888, 181). Here the noun fundr is substituted with the participle fundinn, which corresponds to the noun Svalbarði in nominative case. There is a masculine gender noun barði which is a derivative of barð, although it does not seem to retain its landscape connotations.
Finnur Jónsson (1926, 80) was inclined to assume that Svalbarði was the original ("den riktige") form precisely because it was more unusual; it could have been more easily misread as Svalbarð than vice versa. Although not especially convincing (the earliest manuscript with *Svalbarð predates the Skálholts annáll by about half of a century), the argument was valuable for the tong-in-cheek sabotage of the Norwegian claim to the name Svalbard which we can feel in Finnur Jónsson's article: the Norwegians could appropriate this nice Old Norse name for their new territory, but, first, it did not mean what they thought it meant, and, second, it sounded different from how they thought it sounded. Finnur Jónsson made his point by giving his article the title "Svalbarde" ("Man måtte således også i norsk snarere optage navnet Svalbarde end Svalbard", 81).

22
The same form Svalbarði (or its Norwegianized/Danicized variant, Sualberde) appears in all other medieval sources related to the discovery of 1194 that I am familiar with, except for the Icelandic Annals and recension A of the Guðmundar saga. However, I was not able to find a place name Svalbarði in Iceland. There is a similarly constructed compound Klasbarði, the name of an abandoned coastal farm in Rangárvallasýsla at 63°41'00"N 20°30'00" W; the original medieval nominative form seems to be Klasbarð, while the masculine derivative Klasbarði is assumed to have appeared only in the fifteenth century (Hannes Þorsteinsson 1923, 17).5 This may be an indication that Svalbarði was likewise a chronologically later development.

Medieval Sailing Directions and Willem Barentsz
There is one text that may provide information about the location of Svalbard based on the actual experience of its discoverers. The text includes sailing directions and distances between several points in Norway, Iceland, Greenland, and Ireland. Versions of this text are included in different redactions of the "Book of the Settlement of Iceland" (Landnámabók). The history of the Landnámabók is complex, and its oldest redaction may go back to the beginning of the twelfth century and thus predate the discovery of Svalbard in 1194 (Jakob Benediktsson 1993, 374). The text with sailing directions had been featured in the Landnámabók by at least the second half of the thirteenth century, when Sturla Þórðarson (d. 1284) composed his redaction of the Landnámabók, known as the Sturlubók. The version is preserved in a copy made by Jón Erlendsson in the period 1640-1660 (Stofnun Árna Magnússonar í íslenskum fraeðum, AM 107 fol). Another early redaction, the Hauksbók, was written by Haukr Erlendsson (d. 1334) in 1302-1310; the original manuscript (AM 371, 4to) is incomplete and lacks the sailing directions; a copy of this manuscript was made by the same Jón Erlendsson around 1650-1660 (AM 105 fol).6 There are other redactions of the Landnámabók which are important for reconstructing its history, but which do not provide any significant variants of the sailing directions.
The text with sailing directions was copied into the "Longest (or Greatest, en mesta) Saga of Óláfr Tryggvason" (Óláfs saga 1958, 1: 255-256). The three earliest manuscripts7 date from the late fourteenth century, and it is assumed that the saga was compiled in the early part of the same century (Ólafur Halldórsson 1993, 449). The sailing directions in the saga are based on the Sturlubók redaction and offer no significant variant readings.
Finally, another version of the sailing directions introduces a description of Greenland, which was based on information related by Ívar Bárðarson (Ivar Bårdsøn), likely when he had already returned to Bergen by 1364 after "many years" of service as the episcopal vicar in Garðar in Greenland. This description was first composed in Norwegian, but it survives in a Danicized version which goes back to the early sixteenth century. In his critical edition, Finnur Jónsson listed fourteen manuscripts with complete and partial 5 The origin of Járnbarði as the name of lava cliffs surrounding Dritvík Bay at the end of the Snaefellsnes Peninsula (and derivatives of this name such as Norðurbarði and Suðurbarði, see Brynjúlf Jónsson 1900, 22), could be different and related to (járn)barði, the designation of a combat ship with an iron stem (Hødnebø 1972, 44; see also Þórhallur Vilmundarson 1991, lxxxvi-xci). 6 https://handrit.is/is/manuscript/view/AM02-0105 7
Ívar Bárðarson's version of the sailing directions was creatively adapted to his main purpose, which was to provide a description of Greenland. That is why he omitted the description of the route from Iceland to Ireland and added new information to the descriptions of other routes. He did not omit the route to Svalbard (Sualberde), which ought to mean that he considered Svalbard as belonging to Greenland, but as he did not add any substantially new information, it is obvious that this was not a well-travelled route in his time.
The translator into Dutch (from German) was no other than Willem Barentsz, the discoverer of Spitsbergen. His translation, penned in 1594, was further translated into English by William Stere for Henry Hudson in 1608. Having been unable to access the complete chain of sources for Barentsz' autograph manuscript, I cannot speculate as to when exactly the reference to Svalbard was lost in translation and whether it was Barentsz or his predecessor who understood it as Swawelberch (Burger 1928, 227) which name Stere (1906, 163) correctly rendered as Brimstone Mount. In any event, Barentsz knew, and thought about, a corrupted form of the name that would, over three hundred years later, be conferred on his greatest discovery.
The description of the route to Svalbard in all the main versions of the sailing directions is essentially the same (Íslendingabók. Landnámabók 1968, 1: 33-35, Ívar Bárðarson 1930. The route starts in Langanes in northeastern Iceland (which, we may recall, was across the fjord from one of the more ancient Icelandic Svalbarðs). The versions in the Sturlubók and the Hauksbók (but not in Ívar Bárðarson's description of Greenland) specify the direction to Svalbard as "north". The distance is four sailing daegra, usually understood as 12-hour periods; Ívar Bárðarson confirms this understanding with his translation, "two days and two nights" (there were also attempts to interpret daegra as 24hour periods). Trying to pinpoint the destination of this four-daegra journey northward from Langanes, scholars with lifelong experience in sailing in the Arctic found sufficient arguments in favor of Greenland, Jan Mayen and (modern) Svalbard (compare Isachsen 1907, Holm 1926, Tornøe 1944. Distances between other locations named in the sailing directions are not always perfect, which adds uncertainty to each of these wellbuilt hypotheses. A final bit of information about Svalbard locates it in hafsbotn. What is a hafsbotn? Peter Andreas Munch, the first editor of the Historia Norwegie, suggested that the gigantic septentrionalis sinus, described in this text (currently dated to the second half of the twelfth century), was a Latin equivalent of hafsbotn, and that it designated the northernmost part of the sea washing the continuous stretch of land from Norway to Greenland (Symbolae 31-32; this imaginary stretch of land will be the focus of our discussion in the next sections of this article). Following Munch, scholars often assumed that the Arctic Hafsbotn designated the northernmost part of the open sea navigated by the Scandinavian seafarers (Nansen 1911, 291, 391, 414, 463-465, Finnur Jónsson 1926, 86-87, Simek 1990Simek 2016, 94). A recent analysis of geographical terminology in Historia Norwegie removes one of the pillars of this construction, indicating that the septentrionalis sinus should be a large fjord rather than the northernmost part of the sea, and that, "in line with the idea of how Greenland was connected to Europe", this fjord "would have to be between these two areas" (Historia Norwegie 2003, 118, Commentary II.10).
Specialists in Scandinavian and Finnish folklore also regard Hafsbotn as a proper noun rather than a geographical term when they reconstruct archaic cosmological conceptions of a dangerous locality in the farthest north and equate the names Hafsbotn and Trollebotn as they appear in legendary sagas and medieval ballads with the Finnish Pohja/Pohjola: "These names occur in tales of the journeys of heroes to that far north mythical botn; the object of these journeys is usually to release people who are locked up ('innkvervde') there by trolls".9 However, seventeenth-century descriptions of the world talk about hafsbotnar (in the plural) between Greenland and Norway, interpreting this word as a term rather than as a proper noun.10 The word hafsbotn exists outside of the Arctic context, designating any gulf which cuts deep into the land, where the sea ends and it is impossible to proceed further by sea (Fritzner 1973, 1: 688). Good examples of this usage may be found in the itinerary of Abbot Nikulás (d. 1159/1160) who used the term hafsbotn for the Gulf of Antalya (Simek 1990, 483, ln. 131). He also located Venice in a hafsbotn; elsewhere he indicated that the Alps reach the sea in the botn of Venice, i feneyiabotnum (Simek 1990, 481, ln. 49;482, ln. 122), which confirms that this particular hafsbotn corresponds to the contemporaneous designation of the whole Adriatic Sea as Gulfus de Venetia.11 In the Ynglinga saga (the first section of his Heimskringla, c. 1230), Snorri Sturluson applied the term hafsbotn to the Black Sea (Svartahaf) in its relation to the Mediterranean (Simek 1990, 426, ln. 4). In all three cases, we are dealing with bodies of water (the Gulf of Antalya, the Adriatic Sea, and the Black Sea) significantly narrower than the main sea (the Mediterranean). This usage was not limited to the Mediterranean basin: the legendary genealogy of the Orkney earls, known as Fundinn Noregr, in the beginning of the Orkneyinga saga (c. 1200), mention a "hafsbotn that stretches towards the White Sea and that we call the Gulf of Bothnia".12 In principle, the text of Landnámabók does not preclude our understanding of hafsbotn as a term with a similar meaning. For example, a good candidate could be Scoresby Sound (first suggested by Magnusen and Rafn 1845, 212-213, soon after its exploration by William Scoresby). In modern Svalbard, coastlines provide various opportunities for interpreting hafsbotn as a large fjord, but to my knowledge, such an interpretation has never been offered. 9 Lid 1942, 54 [163], see also Lid 1943, 291, 294 [169, 173];Siikala 1999;Drobin och Keinänen 2001, 159. Unfortunately, these scholars did not provide a list of the legendary sagas which used the name Hafsbotn as the destination of their heroes. 10 Simek 1990, 475, ln. 64, 588, ln. 9; the translation "eine Meeresbucht" on p. 477 should be substituted with "Meeresbuchten". See also maps by Jón Guðmundsson discussed below. 11 See, for example, sources related to the Third Crusade, specifically in reference to events of 1191: Gesta regis Ricardi 1867, 205; Chronica Magistri Rogeri de Houedene 1870, 166. 12 hafsbotn þann, er gengr til móts við Gandvík; þat kǫllu vér Helsingjabotn (Orkneyinga saga 1965, 3).

Medieval Geographical Descriptions and Samsons saga fagra
Scandinavian geographical treatises of the twelfth-thirteenth centuries described a continuous uninhabited stretch of land from Bjarmaland (which they located northeast of Finnmark) to Greenland (Melnikova 1986, 77, 79, 87, 89;Simek 1990, 431, 433, 435, 505-507). They mentioned neither Svalbard nor its hafsbotn, but the earliest such description was composed between 1170 and 1190 (Melnikova 1986, 73) and thus may predate the discovery of Svalbard. This image of Greenland as having a connection to the Eurasian mainland entered Western European cartographical and geographical tradition with a map made in 1427 by a Dane, Claudius Clavus, which Cardinal Guillaume Fillastre included in his edition of the Geography of Claudius Ptolemy. As the first "modern" map added to Ptolemy's atlas, it broke through the limits of his world and exercised a long and lasting influence on Western European ideas about the geography of Northern Europe and the Arctic (see Gautier Dalché 2009, 173-176, 414-415).
An attempt to locate Svalbard in relation to this stretch of land survives in the fourteenth-century "Saga of Samson the Fair" (Samsons saga fagra). The two earlies manuscripts, Stofnun Árna Magnússonar í íslenskum fraeðum, AM 343 a 4to, f. 5v-14r and AM 589 b 4to, are dated to the second half of the fifteenth century. The saga consists of two distinct parts13. The first one is an Icelandic romance created under the influence of the Arthurian cycle, while the second part includes legendary material which is more dependent on Scandinavian mythology. This second part focuses on the stories of King Godmundr of Glaesis vaellir and his son, King Sigurðr of Iotunheimar and describes the region in which these kingdoms are situated. Northeast of Austuruegr (which are lands east of the Baltic Sea) the saga names Risaland; further northeast is Iotunheimar, followed by Sualbardi, and Graenlanz obygder ("Greenland's desert"). The kingdom of Glaesis vellir lies east of Risaland and south of Iotunheimar (Samsons saga fagra 1953, 1: 31-32).
While the part of Greenland which borders Svalbard is uninhabited, Svalbard itself is depicted as populated by unusual kinds of people, two of which are remarkable because of their lifespans. Members of one tribe live to the age of 200, but they seldom have many children. In another tribe called Smámeyjar ("Little Maidens"), which inhabits a cape that protrudes into the sea, people live no longer than 15 years and have children at the age of 7. The people of yet another tribe are distinguished by their "foolish nature", although they can still be considered human. They are further described as having "a mountainous mind" or "the mind of mountain-dwellers" (fialla mannuit). According to Simek whose translations I have quoted, this description "may go back to a joke which can no longer be understood" (1986,259).
Some of the place names in the geographical description of Samsons saga fagra designate mythological localities. Both Risaland and Iotunheimar are lands of giants and trolls who, although mostly associated with the North and East, could be also found anywhere in unfamiliar places (Ármann Jakobsson 2006)14. The blessed land of Glaesis vellir is halfway between our world and the otherworld of the immortals (Egeler 2015, 13 See an analysis of its structure and sources in Zwei Rittersagas 1982, 19-34, 37-39, 135-138. 14 The notion of Risaland could have accommodated the similar-sounding name of Russia (Simek 1986, 255-256), although the geography of the saga also included Russia proper (Ruzsaland, Ruzialand) where the 150-year-old Sigurðr hoped to marry a daughter of the "Russian Jarl" but found his death instead.

Svalbarðs fundr. The Place Name Svalbard and Its Connotations in Medieval and Modern Literature and Cartography
26 40). The saga aligns them all in the distant Arctic but sets aside the farthest periphery of the known world for the real, although hardly accessible lands of Svalbard and Greenland.
Simek (1986, 259-262; 1990, 353-356) demonstrated that the marvelous people of Svalbard with their remarkable lifespans could have been borrowed from an Old Norse list of fantastical peoples, which itself goes back to a Latin source, probably the Etymologies of Isidore of Seville. A version of this list survived as a part of the Hauksbók. The combination of the names of Sualbardi and Graenlanz obygder can likewise point to the Hauksbók as a possible source. Only the Hauksbók version of the sailing directions includes a distance from Kolbensey (an island north of Iceland) "til óbyggða á Groenalandi", which is added right after the distance from Langanes to Svalbard (Íslendingabók. Landnámabók 1968, 1: 34).

Seventeenth-Century Icelandic Cartography
By the mid-seventeenth century Jón Guðmundsson further developed the image of Svalbard by fixing it on his map approximately at the location indicated by Samsons saga fagra. According to his biographer Halldór Hermannsson, Jón (1574-1658) was widely read and well-traveled. Yet he never had any formal schooling and was credulous and uncritical. His "Natural History of Iceland" (Um Íslands aðskiljanlegar náttúrur), written sometime during the years 1640-44 (Halldór Hermannsson 1924, demonstrated a good knowledge of the Hauksbók and sagas and also made use of geographical discoveries by the Dutch, English, and Danes in the northern seas.
Around the same time, Jón made a map of the northern regions. Jón was not the first modern Icelandic cartographer to enhance Western cartographical tradition with elements of Icelandic lore. He may have used the work of predecessors such as Guðbrandur Thorláksson (1542Thorláksson ( -1627, Bishop of Hólar from 1571, and Sigurður Stefánsson, who became schoolmaster at Skálholt shortly before his untimely death in 1594 at the age of 25. The work of Jón and other Icelandic cartographers during the Age of Discoveries is an important historical source on the geographical ideas of their time. It does not, however, appear to be methodologically sound to use their maps as evidence of the actual location of medieval Svalbard (compare Starkov 1998, Derzhavin 2016. Jón Guðmundsson's original map did not survive. Halldór Hermannsson knew at least five handwritten copies from the seventeenth century, which were "either very poorly and indistinctly drawn,15 … or elaborately retouched.16 … Therefore, it is difficult to say how the original looked and what names it included" (1924, xxvii-xxviii). Svalbard appears on two of the "elaborately retouched" maps, the images of which I was able to consult, AM 364 fol (Svalbarden) and GKS 2881 4to (Sualbarde) (see fig. 1). The maps show the coast of Greenland stretching both west and north of Iceland and making an almost 90degree turn in the northwest. Greenland does not appear contiguous with Northern Eurasia, being separated from Bjarmaland and the rest of Eurasia by a strait, Dumbshaf/Dubshaf17. The inscription Svalbarden/Sualbarde is northeast of Iceland in the 15 Det kongelige Bibliotek, GKS 2877 4to, f. 16, 18; "one of the maps" reproduced in Halldór Hermannsson 1924, plate 1; the same map in: Haraldur Sigurðsson 1978, 86. 16 Københavns Universitet, Den Arnamagnaeanske samling AM 364 fol, f. 18r: https://image.landsbokasafn.is/source/AM_364_fol/AM_364_fol.,_0018r_-_33-hq.pdf The copy provides the date of the map as "1640". For two other copies, see footnote 20. 17 Compare Dumbshaf in legendary sagas: "The waters west of Gandvík [the White Sea], north of Norway, were called Dumbshaf. This could originally have meant 'the foggy sea', but by the 14th century authors sea off the coast of Greenland, while Riseland/Risaland is in Greenland's northern interior. In their northwestern corners, the maps show one narrow and two wide gulfs, named Hafsbonderne. This is a Danicized version of the nominative plural of the Icelandic word hafsbotn (the copies of the map were made in Copenhagen in Latin and Danish). The copy in GKS 2881 4to reads "Hafsbonderne, that is the end of the sea" (Hafsbonderne deter Ende paa hafvet), which is a perfect explanation of the meaning of the word hafsbotn, the end of the sea, where further sailing is not possible. The maps show two islands, AEis øe and AEisland in the sea northeast of Iceland, closer to Iceland than to Svalbard. Information in the "Natural History of Iceland" adds more details to the cartographic image. The text devotes considerable space to a description of the attractive natural resources of AEgisland. The enigmatic name AEgisland (AEisland) is known only from this text and from the maps of Jón and his followers. As Halldór Hermannsson suggested (1924,32), AEgisland was Spitsbergen; from English, Dutch, and Danish navigators in the Arctic, Jón collected names synonymous with AEgisland -Prestey and Egerlandt/Egerland. Another place name associated with AEgisland was Spitzbergs óbygda, "Spitsberg's deserts". The ideas conveyed in the text are difficult to understand: "Mountains and high rocks of the land [that is, of AEgisland] are facing that deep ocean, between the deserts of Spitsberg, which is Nordhvalaland [Bowhead-Whale-Land], therefore the ancients called it18 Svalbard".19 According to Halldór Hermannsson (1924,31), Jón used the name of Spitsbergen "without any idea of what it really meant". Nevertheless, although Jón learned about the discovery of Spitsbergen from hearsay, he appears to have been the first to use the ancient concept of Svalbard to make sense of this discovery.
Jón's map impressed Thórður Thorláksson (1637-1697), who had received an excellent education in Copenhagen and Wittenberg, where he published a dissertation with a description of Iceland in 1666. A great-grandson of the pioneer cartographer of Iceland, Guðbrandur Thorláksson, he was a talented cartographer in his own right. In 1668, he compiled a work on Greenland in which he included copies of maps by his Icelandic predecessors (Guðbrandur Thorláksson, Sigurður Stefánsson, and Jón Guðmundsson) as well as his own original map. The work survives in two copies.20 He also made a large paper chart (66x37 cm), dedicated to Henrik Bjelke, governor-general of Iceland in 1648-1683.21 agree that it is named after a king Dumbr, who reigned over this part of the world in times long past" (Simek 1986, 257). 18 Does this "it" refer to the northern coast of AEgisland? Or does it indicate the area of the ocean north of AEgisland, like on the map? 19 Fiöllin og hábiörg landsins horfa fram j þad djúpa mejginhaf, j millum Spitzbergs óbygda, sem er Nordhvalaland, þvi kölludu þeir gömlu þad Svalbard (Halldór Hermannsson 1924, 4, ln. 14-16). Elsewhere (ln. 1-3) AEgisland is also named together, and probably equated, with Svalbard j Hafsbotni, but the text may be corrupt. 20 Det kongelige Bibliotek, GKS 997 fol, written in 1669, Jón's map dated "1640", Thórður's "1669", see: Kålund 1900, 7, Nr 12

28
Thórður Thorláksson demonstrated a similar vision of the sea and the coasts north of Iceland. He confirmed that Hafsbonderne were the end of the sea, as they were always full of drift ice. Sualbarde means a "Cold Side", since this shore was covered with an abundance of ice year-round (Halldór Hermannsson 1926, 40), while the coastal mountain range had snow-covered peaks (43). Only the large chart shows Jón's AEis øe and AEisland in the sea northeast of Iceland, in slightly different forms, AEgirsey and AEgirsland. The large chart differs from both the map in GKS 2881 4to and from copies of Jón's map by placing the inscriptions with definitions of Sualbarda and Trollebotne (instead of Hafsbonderne) on the coast of Greenland rather than on the ocean itself.
Derzhavin (2016) reminds us about the influence of Icelandic scholarly cartography on eighteenth-century Western European maps. Both Jón Guðmundsson's and Thórður Thorláksson's maps from GKS 2881 4to were indeed copied with minor changes by Þormóður Torfason (1636-1719) who also created his own map, drawn for him by Jacob Rasch (1669-1737). Þormóður Torfason published these maps in his book Gronlandia antiqva (1706, set of maps inserted before p. 21, tab. III, IV, V). One of the (probably accidental) changes he made was to an inscription on Jón Guðmundsson's map, which he translated from Danish into Latin, and for which modern historiography has produced a variety of confused explanations (Krawczyk 1987, 226, Starkov 1998). Instead of "Hafsbonderne, that is the end of the sea" the map now reads "Hafsbonderne, that is the end of Sualbarda" (Hafsbonderne id est, finis Sualbardae), while the Danish word for the sea, Hafvet is left untranslated as a separate inscription ( fig. 2).
Þormóður Torfason's book in its turn influenced Hans Egede (1686-1758), a Danish missionary in Greenland and a great polymath, whose map of Greenland22 was reproduced with some changes in other eighteenth-century maps and atlases.23 Both Þormóður Torfason (21) and Egede (1741, 1-3) considered Greenland to be separate from Spitsbergen and Novaya Zemlya. On Egede's and related maps, variations on the legends Hafsbotner and Svalbardi are written in the same manner as on the Þormóður Torfason/Rasch map, in the sea along the eastern coast of Greenland due north or northeast from Iceland.

Russia or Greenland?
Throughout the eighteenth and nineteenth centuries, interest in the Nordic past was steadily growing both inside and outside of Scandinavia, while the advance of Arctic exploration and improved maps provided more tools and materials for the identification of forgotten discoveries. The first scholars who tried to compare the data on Svalbard with the results of contemporaneous exploration were looking for this land in both Northern Eurasia and in Greenland, at both ends of the mythical land bridge "from Bjarmaland to Greenland".
On the Eurasian side, a most useful tool to reconcile new data with traditional knowledge was "The Russian Atlas" (Atlas Rossiiskoi) published by the Academy of insert between pp. 36 and 37; Nørlund 1944, pl. 41 Sciences in 1745. The atlas gained wide international recognition (Postnikov 1989, 34, 50) and significantly improved understanding of Russia's northern coastline. In an Uppsala dissertation (disputatio pro gradu), presided over by Jacob Fredrik Neikter (1744-1803 and defended by Johan Vilhelm Carlström (1768-1795), maps in this atlas were used to locate precisely all the areas named in the sagas. Neikter24 located Bjarmia to the east of the Northern Dvina. At some uncertain point to the east it bordered Risalandia, a part of which, between the Pinega and the Mezen, was Glesisvall. Neikter saw Jotunheimia as the stretch of coast between the Mezen and the Pechora; east of the Pechora he located Granland, or Granlands Obygder, a part of which, "perhaps the island in the Vaygach Strait, now bearing the name Olenoi Ost, was formerly called Svalbard". Smamöyaland was Yamal Peninsula while Helluland or Hellulands Obygder was Novaya Zemlya (Neikter 1794, 8-9). The island which Neikter equated with Svalbard, Olenoi Ost [rov] ("Reindeer Island") is a small island in the Kara Strait. The only explanation for this strange choice is that Neikter misread the name Vaygach as designating a strait between Novaya Zemlya and the mainland (that is, both the Yugorsky Strait and the Kara Strait), and that he in fact identified Svalbard with the large Vaygach Island, to which he mistakenly applied the legend Olenoi Ost (see: Atlas Rossiiskoi 1745, №14). Neikter planned to continue his work and support his ideas with additional arguments and sources but did not have a chance to correct his mistake. In another dissertation, defended by Lars Daniel Hammarén (1779-1833), Neikter attempted to explain the peculiarities of Scandinavian ideas about the Arctic by a confusion of "two Greenlands", namely Grönlandia (the Greenland proper) and Graenlandia (which he identified with Novaya Zemlya), while Hellulandia this time was the continental region across the strait, which Neikter once again incorrectly named Waigaz. This time Neikter did not mention Svalbard (Neikter 1802, 7-12). Jakob Adlerbeth, the leader of the Swedish romantic Götiska förbundet, quoted the first of the two dissertations at length (Adlerbeth 1813, 82-84) in his substantive article on Bjarmaland. He also named Neikter's main sources, including Samsons saga fagra. The article was widely read not only in Scandinavia, but also in Russia.25 The misread names in the atlas may have prevented the identification of Svalbard with Vaygach from gaining currency.
A less precise localization of Svalbard on the Arctic coast of Eurasia "east of Perm (Bjarmeland)" was also suggested by Skúli Þórðarson Thorlacius (1808, 288).
Other scholars who plotted medieval names onto the modern map assumed that Svalbard could refer to East Greenland. In comparison to the times of Hans Egede, there was much more information about its coast. As Finn Magnusen and Carl Christian Rafn were preparing their volume of medieval references to Greenland, they received updates about ongoing exploration. Their ideas about Svalbard of the Landnámabók and Samsons saga fagra evolved in accordance with these updates. Initially, they placed Svalbarði on an either southern or southeastern stretch of East Greenland's coast (Magnusen and Rafn 1845, 46). Subsequently they moved Svalbarði to northeastern Greenlandmore 24 At that time the texts of dissertations were usually authored by the presiding professor; the process of disputatio is well described in Geete 1907. 25  specifically to the area of the largest fjord system in the world, Scoresby Sound, which had recently been mapped by William Scoresby (212-213), and, more precisely, to the "Liverpool Coast", about which the data were not yet entirely conclusive (Rafn 1845, 846). A substantial defense of the Scorseby Sound hypothesis was published by Gustav Holm in the wake of Norway's renaming of its new territory (1926); this hypothesis is still popular among scholars in Russia (see: Starkov 1998, 5-18, Derzhavin 2016.

Svalbarði on Svalbard
The nineteenth and early twentieth centuries comprised the last great chapter of the European exploration of the world. This was also a time of discovery of the world of the past as it had appeared to medieval and early modern explorers. The first comprehensive collection of Old Norse geographical literature was prepared by Erich Christian Werlauff (Symbolae 1821); it was followed by improved editions and analyses of the main manuscripts containing geographical treatises (see: Melnikova 1986, 16-17). It became apparent that plotting the data of medieval sources on a modern map was not sufficient. One had to be able to imagine the discoverer's mental map, which could be reconstructed with the help of both cartographic sources and narrative descriptions.
The evidence of Icelandic geographical literature and sagas, as well as continental sources such as Historia Norwegie and Konungs skuggsjá, was summarized by Gustav Storm (1890Storm ( , 1893. According to this summary, medieval Scandinavians perceived the Atlantic and Arctic Oceans as a nearly-enclosed sea that they navigated with remarkable success. Knowledge of the northern and western coast of this sea, including Bjarmaland and Greenland, was very sketchy. An outer ocean (úthaf) resembling the archaic ocean of Homer encircled this stretch of land on the other side, like the other peripheral regions of the ecumene. The úthaf was the abode of primordial chaos, about which nothing could be reliably ascertained. The úthaf connected with the Atlantic south of Greenland via a strait, which received the name Ginnungagap, the primordial abyss of Scandinavian mythology. Continuing Storm's work, the Danish historian of science Axel Anthon Bjørnbo created a schematic drawing of the known world as Icelandic geographers of the twelfthfourteenth centuries might have perceived it, with a land connection between Northern Eurasia and Greenland (1909, 231, fig. 6). On this drawing Svalbarði is on the southern coast of this land, washed by the waters of Hafsbotn, which represents the northeastern gulf of the sea. This drawing has been and remains a powerful tool for historians of Old Norse geography (Simek 1986, 248, fig. 1).
The analyses of both Storm and Bjørnbo gave a fresh impetus for the search of the identity of Svalbard. These two scholars, as opposed to some of their followers, did not have any deeply held beliefs about the outcome of such a search. Bjørnbo was interested in finding the location of Svalbard on his sketch of the medieval Scandinavian world, but he neither cared nor speculated about the relation of this place name to any modern geographical reality. Although Storm is usually counted as an early adherent of the idea of Spitsbergen's discovery in 1194, in fact he merely offered Spitsbergen as a non-binding guess in one of his works (Svalbardi could be "for example, Spitsbergen", see Storm 1890, 344-345). In his edition of Historia Norwegie he referred to P.A. Munch's opinion that this land was the coast "discovered by Scoresby" (Monumenta historica Norvegiae 1880, 78-79, quote from Symbolae 1850, 32), and in a later article suggested that Svalbard was "evidently Jan Mayen or Spitsbergen" (Storm 1893, 78).

31
The first person in the nineteenth century to have named Svalbard (Svalbardi) and Spitsbergen in the same breath, or at least in the same paragraph, seems to be the founder of Norwegian geology, Baltazar Mathias Keilhau (see Arlov 1988, 46-49). Long before Storm, relying mostly on information from Samsons saga fagra as rendered in Schøning 1751, Keilhau speculated that "our forefathers" had known or had "reliable assumptions" (sikre Formodninger) about the existence of land in the area of Spitsbergen (Keilhau 1831, 225). Based on this and other pronouncements, Berg (2013, 162-163) argues that already Keihlau recognized Spitsbergen's "potential relevance to Norwegian national interests".
In the twentieth century, the hypothesis about the discovery of Spitsbergen by the Vikings and its identity with Svalbard mostly developed in parallel with and sometimes as a part of the political process of securing Norwegian rights over the island group. Primarily, it was the work not of philologists like Storm, but of patriotic men of action. The most brilliant of these, Fridtjof Nansen, was able to find followers outside of Norway (Hennig 1950, 462). Later adherents of the idea worked within a rather isolated Norwegian tradition,26 becoming more and more certain of their conviction and judgment. In his famous book, Nansen expressed some reservations: "it can be considered likely (kan det ansees som sandsynlig)" that Svalbard was Spitsbergen (1911, 415). Closer to the other end of the tradition, Johannes Kristoffer Tornøe no longer had any reservations: Svalbard "can be nothing else but Spitsbergen, perhaps together with Franz Josef Land and Bear Island" (1944,57).
The existing evidence on the discovery of Svalbard in 1194 and its reception would not let us adopt a position about its identity. However, this evidence offers two important parallel stories. One is the story of a name. The name Svalbard was probably borrowed from an Icelandic farm and evolved into different variations, changing gender and meaning and becoming Swawelberch under the pen of Willem Barentsz and Brimstone Mount in an early English text about the Arctic. Having finally come to designate the archipelago acquired by Norway, it now serves as a reminder of the legacy carried forward by the modern Arctic nation. The second story is that of changing visions of the Arctic, which could accommodate the name and its variations. Medieval Svalbard has been sought along the real coastline of Russia and Greenland, but also along an imaginary stretch of land, which connected Northern Eurasia with Greenland on the mental map of medieval seafarers and geographers. Like many other distant lands on the periphery of the known world, this Svalbard has remained elusive, lingering in the realms of geographic myth.