The early Viking ship types

ELDAR HEIDE The early Viking ship types . The Oseberg ship at the Viking Ship Museum in Oslo. (Photo: BSJ) Osebergskipet på Vikingskipshuset i Oslo. (Foto: BSJ) Introduction The Viking ship, a world-famous icon, is known to us from ship finds, iconography, and written accounts, especially Old Norse sagas. Most of the research, however, has focused on the material remains. There is reason to believe that a more extensive utilization of the textual evidence can substantially increase our understanding. For example, we need the texts if we want to know not only what the ships were like, but also what people called them. In this article the author attempts to sift out the early Viking Age terms for ship types from Old Norse (ON) written sources and link them to actual ships and ship depictions from that period. The author argues that knorr, beit, skeið, kjóll, askr, and elliði were the main ship types of the early Viking Age in Scandinavia, at least in the west, and that knerrir referred during this period to warships like Oseberg and only later to cargo ships like Skuldelev . ‘A ship with a backwards curved stem’ seems to have been the original meaning of knorr. Kjólar were heavy, all-round ships like Gokstad, and beit were very early ships with angular stems known from depictions. Skeiðar were low, narrow ships like Ladby. Askar were also very early, small, light ships with stitched planking, whereas elliðar were combined inland / sea vessels, originally Eastern European. Most introductions to Old Scandinavian ship types, such as knorr, snekkja, or karfi etc., are problematic and the majority of the above-mentioned terms are prob- 82 E H ably unfamiliar even to most Viking ship enthusiasts. In fact, much of the literature on Old Scandinavian ships still understands High Medieval terms as referring to (early) Viking Age ships, despite the fact that maritime technology underwent a radical change in the intervening period, developing from fleets of many swift, small craft suitable for beach landings into fewer, much larger ‘floating castles’ for sea battles. The shortcomings of existing literature, however, do not necessarily reflect poor scholarship, but a lack of in-depth research. Interest in Viking ships has been considerable from the nineteenth century onwards and, since the s, groups of scholars have conducted fulltime archaeological and ethnographical research into Viking ships; but thus far, there still has not been any dedicated, systematic attempt to cross-reference the ship types that we know from Scandinavian medieval texts (runic inscriptions and manuscripts) with the finds and depictions of ships that date to this period.¹ I would argue there are several reasons why this has not been done: i. ii. ¹ The information in our written sources is scarce and very scarce in the earliest periods. The texts frequently touch upon maritime aspects, but they hardly ever mention anything resembling definitions. We always have to infer indirectly from information supplied for other reasons, even in material relating to the High Middle Ages. Our richest sources – the prose narratives of the High Middle Ages – present ships and shipping from the entire Old Norse period as more or less the same, with terms from across the period frequently mixed together. But the period lasted half a millennium and, during this time, Scandinavian maritime technology underwent I am now attempting to fill this gap in the scholarship through a series of articles. T  V   83 an enormous development. Accordingly, the texts give us a deceptive image of the early ships, which must have been very different from those of the High Middle Ages, which the authors knew. iii. Even if this impression may in principle be corrected by the poetry that probably originated in the Viking Age and was transmitted orally until it was recorded in the High Middle Ages, the information provided by poetic sources is very limited. iv. Both the maritime technology and its terminology were international in nature, with borrowing back and forth, in and out of languages, and this complicates reasoning, especially with regards to etymologies. v. Some of the relevant terms were in use for such long periods of time that their referents changed considerably, causing confusion. vi. Both finds and depictions of ships from the Middle Ages are hard to interpret. The depictions are conventionalised or coarse, while the ships are only partly preserved, which means that defining details may be beyond our reach. vii. The finds of Viking ships are so few that we lack archaeological information about even major ship types. viii. In order to combine the written material, linguistic and etymological reasoning with the archaeological, iconographic and ethnographic material, it is necessary to take an interdisciplinary approach and so combine a level of practical and scholarly knowledge that is very hard for any person to attain. As is customary, I will use the term ‘warship’ even when ‘all-round ship’ would often be more precise. Before the development of specialised cargo ships (probably during 84 E H the course of the tenth century), cargo was transported on ships that were also used for warfare and which we therefore tend to call ‘warships’. I exclude small boats from the discussion because the topic of ships and large vessels is more than difficult enough on its own, and because it seems possible to separate boats from ships in the material. Two terms are excluded because they refer to small vessels, eikja and nokkvi (§  here), which designate boats for only one or two persons. Even the Viking ‘ships’ were often quite small by modern standards. . Identification of the early Viking Age terms The key problem is described in point ii. in the above list. If we fail to decode the mixing together of information from diverse periods within the five centuries in question, we will never be able to correlate the ship-type terms with actual ship types revealed through material finds and depictions. The different periods used different types of ships referred to by terms that in most cases underwent significant changes. Hjalmar Falk () was unaware of this problem and was consequently unable to say much about the developments across the Old Norse period (late eight to the fourteenth century) or about differences between various parts of the Old Norse period. Rikke Malmros (, reprinted in ) pointed out that the Viking Age skaldic poetry gives an impression of ship types different to the one given by the High Medieval prose in which the poetry has come down to us, and Judith Jesch (b) has taken this approach a long way forward by comprehensively examining all the late Viking Age skaldic poetry and runic inscriptions. Narve Bjørgo () has examined the material at the opposite end of the time-scale, namely the contemporary (High Medieval) kings’ sagas, and presented a specific picture from that period. Nonetheless, much remains to be done in terms of categorising the Old Norse ship terms by T  V   85 period. For example, Detlev Ellmers, Kim Hjardar and Vegard Vike are still mixing terms and information from across the Old Norse period.² In this article, I attempt to look at terms from the early Viking Age – here defined as the period from the late eighth century to the mid-tenth century. The contemporary source material from Scandinavia is limited to a few brief runic inscriptions. Still, there is material that we can use. I will base my arguments on the following types of primary sources: a. Terms mentioned in runic inscriptions from the early Viking Age. b. Terms found as loanwords or cognates in Northern European countries that have written traditions earlier than Scandinavia and that were recorded during or before the early Viking Age. c. Bynames of people believed to have lived in the early Viking Age. d. Terms occurring in skaldic poems that date from the early Viking Age. e. Terms occurring in Eddic poetry. f. Terms that occur only in Old Norse poetry and are not found in prose. This combination of source types, and source type e. and f., are my ideas. The material compiled in order to investigate points a., d., e., and f. can be found in the appendix (§ ), together with information on how it was selected. Contemporary Viking Age information written down in other Northern European countries (b.) probably has quite a significant information value – especially when the ship type terms are borrowed from Scandinavian and/or refer exclusively to Scandinavian ships, but in other cases, too, because it seems that the maritime technology was international in nature and ² Ellmers : –; Hjardar and Vike : –. 86 E H spread rapidly between countries, in the Viking Age as in later periods. The bynames (c.) are handed down to us in manuscripts several hundred years younger than our period of study. But the bynames in question were borne by some of the first settlers in Iceland or their children, which means that they form part of the essential memory of the founding of the Icelandic nation, and therefore they can be taken as quite reliable. Source type d. is dependent on Finnur Jónsson’s dating based upon the High Medieval tradition, which is obviously problematic. However, there is broad agreement, even among strictly source-critical scholars such as Claus Krag and Annette Lassen, that most of the allegedly early skaldic poetry is authentic, and there is now a trend to verify the outlines of Finnur’s dating.³ For my purposes, it is advantageous that Finnur rejected many poems – especially from the sagas of Icelanders and the legendary sagas – as ‘fake’ because this reduces the risk of mistaking late terms as early ones. As source material, skaldic poetry also creates problems due to its heiti system of ‘synonyms’, which allows any term belonging to a semantic category to be replaced by another. Because of this, it is hard to know whether a skald really meant a knorr or a skeið, for instance, or whether these terms just represent the category ‘ship’. Still, if a term was used, it was apparently known and in the cases where a ship-type term occurs with a characterising adjective we can assume that the ship type really was intended, because if not then the characterisation would miss its mark. Source type e. is very uncertain as Eddic poems are even harder to date than skaldic verse. Still, most of these poems are widely understood as pre-Christian, i.e. from before or around  AD, although this is impossible to know and the forms handed down to us from High Medieval manuscripts may deviate from the pre-Christian ³ E.g. Bergsveinn Birgisson  and Myrvoll . For Krag and Lassen, see e.g. Krag :  and Lassen : . T  V   87 forms.⁴ I do not emphasise evidence from Eddic poetry except in the discussion of the term kjóll (§ ). Source type f. is based upon the fact that poetic words are often survivors from earlier everyday language (e.g. the Modern Icelandic poetic terms ver m. and höldur m., ‘a man’, and mögur m ‘son’, all of which were common prose words in Old Norse). Therefore, given the size of the Old Norse prose corpus and its extensive references to ships and sailing, ship-type terms found in poetry, but not in prose, probably belong to the earlier stages of the Old Norse language. This assumption seems to be confirmed by the cases of askr, elliði, and kjóll. The former two ship types are mentioned many times in poetry, but never in prose – with the exception of a few occurrences in the late legendary fornaldarsogur, where the narratives take place before the settlement of Iceland (see §  and ) and where the use of these terms can most easily be understood as deliberate archaisms. It thus seems that writers in the High Middle Ages understood these ship types as belonging to ancient times and, in both cases, the antiquity of the terms in question is supported by independent material (§ , ). A similar circumstance is found in connection with kjóll, see § . Etymology can also sometimes demonstrate that a certain term existed at some point in the past; but this does not necessarily mean that the term’s referent did, because new referents may be added (see § ., , ). For this reason, I have not included etymology on the list above, although it is useful for some of my arguments. Most of the information that I use to identify the earliest terms for Viking ship types is problematic and some scholars will perhaps argue that only source type a. is acceptable. But such a limitation would yield a less reliable result for the present study since we would then certainly lack major terms. The matching of ship terms and ship types, and the considerations that can help ⁴ See e.g. Meulengracht Sørensen ; Fidjestøl and Haugen ; Sävborg . 88 E H us in that regard, require that all terms for major ship types are included in our list of candidate terms. We should therefore be content with probable information and collect as much of this as possible in order to see how it combines before any potentially valuable primary material is discarded. Many less reliable pieces of information can, when taken together, form a structure that becomes believable if several such pieces independently point in the same direction.⁵ In the discussions below, there are many examples both of this and of the rejection of isolated pieces of information. . The ship-type terms of the early Viking Age The terms gathered with the help of the above criteria are listed in the appendix in § , to which the following discussion will refer, except with regards to etymology, bynames, and the loanwords, which are added in the discussion. The following discussion will refer to this material. In alphabetical order, the terms are askr m., barði m., bátr m., beit f., dreki m., eik f., eikja f., elliði m., ferja f., flaust n., fley n., herskip n., karfi m., kjóll m., knorr m., langskip n., lung n., naðr m., nór m., nokkvi m., regg n., skeið f., skip n., and snekkja f. Certainly not all of these actually belong to the early Viking Age. Firstly, the attestation of karfi in the lausavísa attributed to Egill Skallagrímsson in  is isolated; the next is found in Sigvatr’s Austrfararvísur from ,⁶ and only in High Medieval prose does karfi become common. As Foote points out, the circumstances of the lausavísa attributed to Egill are not convincing.⁷ Finnur Jónsson’s earliest example of snekkja (Guttormr sindri, tenth century) may be false, too, as one manuscript variant reads skeið. Russell Poole prints snekkjum, the reading of his ⁵ ⁶ ⁷ Cf. Heide , Heide . Finnur Jónsson – BI: . Foote : . T  V   89 main manuscript Kx and the other Heimskringla manuscripts, noting that most editors have favoured this reading because it yields a regular internal rhyme (aðalhending) with the second syllable of Eirekr (boðsœkir helt bríkar, / brœðr, sínum, ok flœðu / undan, allar kindir / Eireks, á haf snekkjum).⁸ At the same time he notes that skeiðum is also possible since the rhyme r : ð (Eireks – skeiðum) is permissible and is paralleled in st. / of the same poem.⁹ One could argue that skeið should be preferred because the unusual rhyme (r: ð) makes it the lectio difficilior.¹⁰ This would fit better with the fact that there seems to be no other evidence of snekkja at this early stage. The first quite certain Old Norse example is from  and it is recorded a number of times in contexts from the mid-eleventh century.¹¹. In Germany and England, the related snacgun and snacc are not attested until around  and the mid-eleventh century respectively¹². There are several indications that the ship type snekkja is younger than skeið and that it largely replaced it (as I will argue elsewhere). With regard to the rest of the terms, I see no reason why they cannot all be early Viking Age, although I will only argue that some of them actually are. Many will be excluded because of insufficient evidence. In fact, the evidence only allows us to connect depictions and ship finds with the major terms for specific types of ships. Generic terms are skip ‘a ship’, bátr ‘a boat’, far, literally ‘a means of transport’, and probably flaust and fley, both of which literally mean ‘something floating’¹³. Herskip ⁸ ⁹ ¹⁰ ¹¹ ¹² ¹³ Whaley : –. Kuhn : . Thanks to Diana Whaley and Russell Poole for help with this stanza. A rule in textual criticism is that where different manuscripts conflict on a particular reading, the more unusual one is more likely the original, because scribes would tend to replace odd words and uncommon sayings with more familiar and less controversial ones, rather than vice versa. Finnur Jónsson – BI: ; Jesch b: . Summarium Heinrici : ; Thier : , . It is possible that snac is also mentioned in the late tenth- century Exeter Book, in Riddle  (Williamson in Bitterli : ). Falk : ; de Vries : , ; Ásgeir Blöndal Magnússon : , . 90 E H ‘warship’ and langskip ‘longship’ also covered several types of warships (for example, landvarnarskip / leiðangrsskip, attested later), and there is no evidence linking these terms to the early Viking Age. Much the same may be said of naðr ‘a snake, dragon’, as well as the similar dreki and ormr.¹⁴ The High Medieval sources indicate that these three words were additional terms – if a skeið, for example, was adorned with a dragon head, it was a ‘dragon’, but it still also remained a skeið.¹⁵ We have many depictions of ships with dragon heads from the early Viking Age, but our written material is too sparse to tell us whether the terms naðr, dreki and ormr can be dated so early. Some of the terms occur only once or twice in the appendix material and thus seem not to have referred to major ship types: barði, eik, regg, lung, nór, ferja, and elliði. For some of these, there are additional reasons to omit them from further discussion. Eik, literally meaning ‘oak’ and never referring to a ship in prose, is not necessarily a term for a ship type at all; it may be a metonymical use where the material the ship is made of refers to the ship. Lung is not securely attested until  (Hallfreðr) and is used only a handful of times after that. It is never used in prose and this indicates that it is early, but it seems not to have been among the major types. It may be an early synonym for langskip since it is believed to be an early loanword from Old Irish long ‘a ship’, ¹⁴ Today, we would distinguish fairly sharply between the creatures referred to as ‘dragon’, ‘snake’ and ‘worm’ and would not expect them to be used synonymously. But it seems that our distinction between snake and dragon was not made in Old Norse. Óláfs saga Tryggvasonar in Heimskringla  (I: ) tells of the ship that King Óláfr allegedly confiscated from the chieftain Rauðr hinn rammi just prior to  AD: Þá tók Óláfr konungr dreka, er Rauðr hafði átt …. Þat skip kallaði konungr Orminn, því at þá, er segl var á lopti, skyldi þat vera fyrir vængi drekans: ‘It was a dragon, and the king named it the Snake because when its sail was aloft, it resembled the wing (sic) of the dragon’. Ormr and naðr never occur in prose descriptions of ships with the exception of ormr in the ship names Ormrinn (skammi) and Ormrinn langi; these words seem to be poetic synonyms for dreki. ¹⁵ This is most clearly demonstrated in Hákonar saga Ívarssonar (: ): Lét konungr þá setja upp hofuð þau hin gyltu; mátti þá kalla skipit hvárt er vildi skeið eða dreka: ‘The king then had the golden dragon heads put up; then one could call it skeið or dreki according to one’s wish’. T  V   91 allegedly derived from the Latin nauis longa ‘longship’, but this is uncertain.¹⁶ Regg is also attested exclusively in poetry, but only once (early eleventh century) in addition to in Þulur.¹⁷ Nór is reflected in the name of the god Njorðr’s residence, Nóatún – literally ‘ship farmyard’ – which indicates that it is old, and this can also be seen from its etymological relation to Latin nauis ‘a ship / boat (in general)’.¹⁸ But it is found only twice in the poetic corpus: in the kenning brandnór ‘hearth ship’, ‘house’ § .; and in the Þulur.¹⁹ In prose, nór is only attested twice, both times in the meaning ‘a trough’.²⁰ With regards to ferja ‘a ferry’, etymological cognates in other Germanic languages indicate that the word is ancient, but it is difficult to attest in Scandinavia in the early Viking Age as it occurs only twice in Old Norse poetry.²¹ This may be because ferja was a primitive, low status boat type and the term was therefore unsuitable in skaldic verse, which was highly sophisticated in style. In prose from the High Middle Ages, ferja is very common. – Eikja and nokkvi appear to be boats; there seems to be no Old Scandinavian attestations where they necessarily mean ‘ship’ and, in later times, they refer to small (primitive) boats.²² Both eikja and nokkvi seem to be really small, only fitted with one or two pairs of oars. ¹⁶ Falk : , . ¹⁷ (Nafna)þulur is a subsection of Snorri Sturluson’s Edda, the last part of Skáldskaparmál. It is a listing in verse of terms that may be used in poetry for various items, such as gods, giants, people, animals, and weapons. The þulur are thought to be a later addition to Snorri’s original composition from the s and are therefore sometimes omitted from editions and translations of his Edda. ¹⁸ Simek : ; Falk : . ¹⁹ Finnur Jónsson –: , . ²⁰ http://dataonp.hum.ku.dk/. This is the only meaning it has in Modern Icelandic and the normal meaning in Modern Norwegian, but in eastern Norwegian, the meaning ‘a primitive rowing boat’ existed until the twentieth century (Sigfús Blöndal : ; Aasen : , ; Refsum ). ²¹ Falk and Torp –: ; Finnur Jónsson –: . ²² Fritzner – II: ; Aasen : ; nokkvi: Sigfús Blöndal : ; Nachen: Grimm and Grimm – : . For nokkvi, there is one exception in Edda Snorra Sturlusonar :  and one in Þorbjorn hornklofi’s Glymdrápa, see § . below. The former may be ironic and the latter is found in a corrupt half stanza; Finnur Jónsson –:  92 E H Barði is only mentioned twice in the appendix; and the latest attestation – by Einarr skálaglamm – seems to be copied from the earliest – by Bragi gamli – which may imply that only Bragi’s use of the term has information value for actual ships. On the other hand, Einarr’s allusion to Bragi may be taken as a confirmation; in the late tenth century, Einarr apparently understood Bragi’s poem as evidence that there was an old ship type barði (which is significant as the kenning borðróinn barði is somewhat irregular, because if barði alone means ‘a ship’, the determinant borðróinn would not be necessary to give the meaning ‘a ship’). Finnur Jónsson lists two more attestations of (*)barði, but points out that they may instead be the genitive plural of barð n.,²³ ‘the transition piece between the stem and the keel’, in which case the term is a pars pro toto for ‘a ship (in general)’. The rest of the certain attestations of barði as a ship type seem to date from the late tenth and early eleventh centuries. In Old English, probably at this time, there are a handful of attestations of barð and barða as a term for a ship type, borrowed from Old Norse.²⁴ In Old Norse sources, the attestations of barði at this time all refer to the famous ship (Jarn)barði(nn), owned by Earl Hákon and his son Eiríkr. Both Tindr Hallkellsson around  and Halldórr Ókristni around  refer to this ship in poetry and it is mentioned several times in later prose.²⁵ Even Einarr skálaglamm’s borðróinn barði refers indirectly to Jarnbarðinn (because the earl does not accept him, he threatens to enter the barði of the earl’s rival.) Scholars understand Bragi’s barði and the earls’ (Jarn) barði(nn) as examples of the same ship type, but this is problematic. In both cases, barði apparently derives from the common noun barð, but hardly in the same way. There are  years between them and, according to ²³ Finnur Jónsson –: . ²⁴ Pers. comm. from Katrin Thier ..; Thier : . ²⁵ Finnur Jónsson – BI: , ; Falk : . Homlubarði in Konungs skuggsjá in all probability does not refer to a ship but to ‘rocky ground’, see Schnall . T  V   93 High Medieval tradition, the ship (Jarn)barði(nn) was so named because it was reinforced with bands of iron across the borð (pl.) and other parts of the bows. This clearly relates to technological innovations of the late tenth century (which I will discuss in a forthcoming article), so this can hardly be the background of Bragi’s early or mid-ninth century ship. Bragi’s use of barði is therefore left quite isolated, a fact indicating that it was not a major ship type. But see § . even so. Assuming that the reasoning here is correct, we are left with askr, beit, elliði, kjóll, knorr, and skeið (which are therefore given in bold type in § ). Kjóll, knorr, and skeið seem to have been major ship types because they are mentioned many times in the appendix material. Askr, beit, and elliði are not, but regarding these, there is additional material indicating that they are significant and early ship types (see below). Accordingly, these are the terms that we should attempt to connect with the finds and depictions of ships from the early Viking Age – and this will be the main focus in the rest of the article. The terms will be discussed in the order that I expect will make the reasoning easiest to follow. . Ship characteristics, terms, etymology, and stem profiles When attempting to identify terms for ship types with actual ships and depictions, it is important to try to uncover the original meaning of the terms – their etymology. However, these etymologies need to connect to traits that seem to have characterised the ship types in question (during the Viking Age or at an earlier stage). Many of the etymologies suggested are either so general that they would fit any ship, or refer to characteristics that the ship types in question can hardly have had, see for example askr, skeið, and kjóll in § ,  and . Terms for ship types may refer to all kinds of characteristics, as can be illustrated through some transpar- 94 E H ent terms for boat types known from eighteenth to nineteenth-century Scandinavia, particularly the west coast of Norway: - Characteristics of hull: snidbetning ‘boat with aslant planks in the bottom’, veng(je)båt ‘boat with a veng (a particular kind of cabin)’, gavlbåt ‘boat with a gavl (blunt rear)’.²⁶ - Building materials: e(i)ka ‘(boat) of oak’, äsping ‘(boat) of aspen’.²⁷ - Boat usage: skötbåt ‘boat for net (sköt) fishing’, straumbåt ‘boat for the Saltstraumen tidal current’, lofotbåt ‘boat used by the Lofoten fisheries’.²⁸ - Cargo capacity: lestabåt ‘a boat that can hold twelve barrels’, bunkerømming ‘a boat with a bunkerom, or extra space for cargo’.²⁹ - Size: storbåt ‘big boat’, sjun ‘the smallest type of Nordland boat’, literally: ‘something barely visible’ (= Old Norse sjón f. ‘sight, vision’).³⁰ - Vessel size expressed by the number of strakes (Old Norse borð n.): tribør(d)ing, firbør(d)ing, fembør(d)ing.³¹ - Vessel size expressed by the number of oars: (Old Norse) feræringr, sexæringr etc. (> Modern Norwegian færing, seks[æ]ring). - Area of origin: nordlandsbåt, åfjordsbåt, oselvar, strandebarmar, tanabask, etc.³² - Stem profile. See below. ²⁶ ²⁷ ²⁸ ²⁹ ³⁰ ³¹ Færøyvik : ,  ff., –. Falk : ; Korhonen . Ahlbäck and Bonns ; Klepp : ; Godal and Eldjarn : ; Klepp : , , . Eldjarn , Fritzner – III: ; Aasen : . Aasen : , , . Hustad and Klepp  (cf. Klepp : ) argue that fembør(d)ing derives from a hypothetic Old Norse *fimbyrðingr ‘a swift, small cargo ship’, but this is improbable when we take into account the parallel terms tribør(d)ing and firbør(d)ing and the fact that fembør(d) ingar with five strakes are known from eighteenth and nineteenth-century drawings and photographs (Eldjarn : –). ³² Færøyvik , Klepp . T  V   95 It is important to realise that any type of craft may be referred to using alternative terms. For example, a straumbåt (use) is also a nordlandsbåt (area of origin) and (usually) a færing (number of oars). A tribør(d)ing (number of strakes) may be an oselvar or a strandebarmar (area of origin) and at the same time for example a færing or a seksæring (number of oars). A snidbetning (hull characteristic) is often synonymous with sunnmørsbåt (area of origin) and at the same time for example a færing or an åttring (number of oars). Such variation is not specific to boat terms: in much the same way, back-country skis (Norwegian fjellski) are also wooden skis (treski) or fibreglass skis (glasfiberski), while an SUV can also be an off-roader and either a diesel or a petrol car. It is also important to bear in mind that designs change over time, in many cases substantially, so a term’s connection to the original characteristic may be lost. For example, a fembør(d)ing has had more than five strakes for a long time, while a tendring, from Old Norse teinæringr ‘a boat with  oars’ (probably borrowed from Low German), is in modern times a sailing craft with hardly any oar propulsion at all. Such changes are a major problem to the study of Old Norse and other early terms for ship types. Ship graffiti from the Viking Age give important information of ship characteristics from the period; more important, I believe, than has been realised. They are far from unproblematic because they may be conventionalised or inaccurate. However, the alternative is ship finds, which with few exceptions lack the stems and profiles that seem to have been important in the naming of the different ship types (see below). I will therefore start with the graffiti and attempt to identify some major ship types from them. These graffiti show a strong interest in ships’ stems and so do other graffiti from the Middle Ages.³³ A similarly keen interest in stems can also be identified in Old English poetry, especially in ³³ E.g. Blindheim . 96 E H Figure  Figure  Figure  Figure  Figure  Figures –: Graffiti from the Oseberg find, between  and  AD (some have been rotated to make the gunwale or keel horizontal). (Christensen : ) Figure  Figur –: Graffiti frå Oseberg-funnet, mellom  og . (Christensen : ) T  V   97 the late eighth-century poem Andreas.³⁴ I believe that the reason for this interest is that the profile of the stem and bow characterised ship types to a much higher degree in the Viking Age than in later periods. From the late Middle Ages onwards, many ship types were distinguished by the number and the type of their masts and sails, e.g. barque, schooner, and brig. Before this there was only one type of mast and sail and only one mast and sail on each ship. This implies that the profile of the stem was one of the few traits that made it possible to identify ships from a distance. As should be expected from this, we can see in the sources that stem types and stem profiles really did serve to categorise ships. A good example is a Norwegian diploma from . It records that the farmers of Sande in Eastern Norway are going to build a new levy (leiðangr) ship and that the king gives them the choice of whether to make it buttu stæmfnt or holka stæmfnt ‘with a stem like a búza (²/bu:tsa/) or with a stem like a holkr’.³⁵ This implies that people at the time conceived of two major categories of ships distinguished by their stems – although not only by their stem profiles, there were also fundamental differences in construction.³⁶ Stem profiles are also reflected in several of the Old Scandinavian terms for ship types. The clearest example is the High Medieval term skúta f., which refers to some kind of small all-round ship. The term is closely related to the verb skúta ‘to jut’ and to the masculine skúti ‘a jutting rock’, which seems to imply that a skúta had a jutting stem, like Skuldelev  – which seems reasonable because small ships more often than bigger ships had projecting stems.³⁷ The term knorr also seems to be a quite certain example of a term reflecting a stem profile (§ .). ³⁴ Krapp , e.g. line . ³⁵ Diplomatarium Norvegicum – II: ; cf. Norges gamle Love – V: ; Falk : ; Hødnebø : . ³⁶ See Greenhill . ³⁷ see e.g. the Bayeux tapestry; Heide : ; Falk : ; and § . here. 98 E H Figure  Figure  Figure  Figure  Of the depictions shown above, we can distinguish between three main types based on the stems and bows of the ships: Figures –. Details from the Skomrak beater in Figure  (Some have been rotated or flipped). A. Ships with backward curved stems. B. Ships with curved stems that project; some with a horizontal top. C. The same as in B – and a triangle ‘filling’ the space below the curve of the stem, making the stem appear straight and vertical. Figur –: Detaljar av vevsverdet frå Skomrak som er vist i figur . The same three profiles – and not others, it seems – can also be found in other depictions and ship finds from the Viking Age (the depictions on the Skomrak beater could be (small) boats rather than ships, but as there are other depictions as well this makes little difference).³⁸ I will begin the discussion by arguing for a connection between stem-type A and the term knorr, and between stem-type C and the term beit. The remaining terms seem to derive from other traits of the ships than the stem profiles. With regards to stem-type B., there are no particular reasons to believe that a certain term ³⁸ Lindqvist –; Varenius ; Felbo . T  V   Figure : Whale-bone weaving beater from Skomrak, Vest-Agder, southern Norway, early Viking Age. (Gjessing : ) Figur : Vevsverd frå Skomrak i Vest-Agder, tidleg vikingtid. (Gjessing : ) 99 was derived from it, but it probably belonged to the traits that defined some of the types. This is because the major term knorr’s very probable connection with a backward-curved stem only makes sense as a contrast to other stem profiles and there seems to have been an awareness of this connection throughout the Middle Ages, even into modern times in local traditions. . Kno˛rr Evidence of at least two kinds indicates that the term knorr m. (pl. knerrir) existed in the ninth century. Firstly, knorr is attested in Þorbjorn hornklofi’s Haraldskvæði  (§ .), which is widely believed to be genuine and composed shortly after the Battle of Hafrsfjord in the s, which the poem describes.³⁹ The occurrence is supported by alliteration – knerrir kómu austan / kapps of lystir – but not by internal rhyme. Secondly, cnear is attested (twice) in Old English soon after , in the Anglo-Saxon Chronicle, Battle of Brunanburh.⁴⁰ If the dating of these occurrences is trustworthy, the term knorr / cnear was established before these dates. In addition, a woman – Þorbjorg Gilsdóttir – whom the sources provide with the byname knarrarbringa ‘knorr chest’ was probably born in the s or s (her father was probably born before  AD; he was among Iceland’s first settlers (see further below)).⁴¹ Regrettably, there is reason to doubt the authenticity of Þorbjorg’s ³⁹ Krag : . ⁴⁰ Two Saxon chronicles : –; Livingston : , , , . ⁴¹ Landnámabók : , , , , and other sources; see Jesch b: . 100 E H byname knarrarbringa (see below), but the mentions of knorr / cnear in Haraldskvæði and The Battle of Brunanburh nonetheless give quite firm ground for an early dating of knorr / cnear. However, in the case of knorr, it is not enough to document an early existence of the term, because it seems to have had two phases with distinctly different meanings. In the High Medieval prose, knorr is common in the meaning ‘(ocean-going) cargo ship’. But Malmros has pointed out that in the poetry composed prior to and around the year , knorr in all clearly defined examples refers to ships used for warfare (cf. § . here).⁴² The warship meaning also seems to fit with the majority of the runic occurrences, of which six exist from eleventh-century Sweden, one possibly from the late tenth century.⁴³ In addition, the warship meaning is the only one in Old English where cnear is mentioned several times in the tenth century. The English meaning seems to reflect Scandinavian usage, because cnear refers to Scandinavian ships exclusively and was in all likelihood borrowed from Old Scandinavian (as can be seen from the form).⁴⁴ Etymology may help us understand the term knorr and the development of the ship type it refers to. Several etymologies have been suggested. One cluster is based upon a root knarr- meaning something akin to ‘a knot in wood, a gnarl’, as in Middle English knarre. Falk argues that knorr ‘referred originally to a gnarled tree-root used as a prow’.⁴⁵ De Vries and Ásgeir Blöndal Magnússon both favour de Vries’ essential idea: The stems were made of knotty wood.⁴⁶ Sayers combines this etymological root with an Old English phrase nægled cnear ‘knorr with planks nailed together’ (attested soon after ) and suggests that ‘knobbed, gnarled’ and the term ⁴² ⁴³ ⁴⁴ ⁴⁵ ⁴⁶ Malmros : . § . here, Jesch a: , Varenius : –, cf. . Bosworth and Toller : ; Thier : . Shetelig and Falk : . De Vries : ; Ásgeir Blöndal Magnússon : . T  V   101 knorr refer to ‘the nail-studded outer hull’ of this kind of ship, because he assumes that a knorr was a cargo ship which would have higher sides and so show more nails than other types of vessel.⁴⁷ Varenius suggests that knorr refers to the squeaking sound of the hull in the sea, apparently taking the Swedish verb knarra ‘to squeak’ as his point of departure.⁴⁸ None of these etymologies, however, are convincing.⁴⁹ When searching for a more plausible etymology, we should have early Viking Age ships used for warfare rather than later cargo ships in mind and we should try to connect them with a root knarr- that is found in connection with ships. Such a root has been mentioned by two scholars, but for some reason little has been made of it. In , Falk mentioned the Norwegian adjective knarrstemnd,⁵⁰ which literally means ‘with a stem like a knorr’, and refers to stems that are backward curved in the stem (‘tilbagebøiet i Stavnen’ – from Helgeland, Northern Norway). This shows what the shape of a knorr’s stem had been like, Falk said, but he ignored this observation in his etymology of knorr. However, Ahlbäck has pointed out (apparently unaware of Falk) that the shape of the stem must be the etymological meaning of the term.⁵¹ Ahlbäck says this in a discussion of the traditional Ostrobothnian term and type knarr / knärr, but the Old Norse term is the same word. There are several reasons to believe that knorr originally meant ‘a ship with a backward-curved stem’.⁵² Firstly, terms that seem to correspond formally to knorr, ⁴⁷ The Battle of Brunanburh, Two Saxon chronicles : –; Livingston : , ; Sayers : –. ⁴⁸ Varenius : . ⁴⁹ For details, see Heide : . ⁵⁰ Falk : –. ⁵¹ Ahlbäck and Bonns : . ⁵² For details, see Heide : . It is unclear how this ‘backward-curved’ meaning of knorr- / knarr- relates to the other meanings. It is conceivable that the different roots knarr- are in fact one and the same; but a semantic connection is not immediately apparent. However, this question does not matter for the present study. 102 E H and that mean ‘with a steep or backwards-tilting neck or stem’ or similar, can be found throughout Scandinavia: - Adjectives: knerren, knerr (Southern Norway), knarr, knärr (Sweden), knarreistur (Iceland) ‘holding one’s head high’, ‘stiff in the neck’; knärrog, knärrnackog (Swedish Ostrobothnia, Finland) ‘with a backwards tilting head’; knerrhava ‘on runners or skis with steep and high tips’; knerrut (Dalarna, Sweden) ‘haughty, with a straight posture’. - Verbs: knärra (nacken) (Dalarna, Sweden; Swedish Ostrobothnia, Finland) ‘turn up one’s nose, be proud or haughty’, knerre (Norway) ’to straighten one’s neck, pull in one’s chin and then raise one’s head’ – and others. - Nouns: knärr (m., Swedish Ostrobothnia, Finland) ‘stiffness in the neck that makes one carry one’s head so it is tilting backwards’; knerre (m., Norway) ’person with a straight posture’; knärrkälke (m., Swedish Ostrobothnia) ‘a sled with steep and high runner-tips’. This set of terms also refers to boats. The Northern Norwegian adjective knarrstemnd has been mentioned above. In Swedish Ostrobothnia, a knärrstam m. is a steep stem on which the upper end sometimes tilts back slightly. If a boat is knärrog (adj.), it is blunt in the bow, with a steep stem. The fact that most of the mentioned words have the vowels e, æ (ä) and a rather than o does not imply a problem. The paradigm of Old Norse knorr includes three different vowels: knorr (nominative singular), knarrar (genitive singular), and knerri (dative singular) or knerrir (nominative plural). The o and e are derived from the a as a result of mutation, knorr from *knarruR, and knerrir from *knarrīR. Thus, the above-mentioned modern Scandinavian words do correspond to Old Norse knorr. The knarr- forms contain T  V   103 the original vowel; the knerr- (knärr-) forms represent the i-mutated vowel; and knorr shows the u-mutated form of the word. If we were to take the word knorr and use it to construct an English adjective ‘*knorry’ or a verb ‘to *knorr’, modern Scandinavian adjectives like knerren / knärrog / knerrut and verbs like knerre / knärra would correspond formally to these. Secondly, modern Scandinavian boat types termed knorr / knarr / knärr, in Northern Norway and Swedish Ostrobothnia, have steep or backwards-curved stems and they contrast with boat types that have projecting stems. The Northern Norwegian form is knorr (knørr in some dialects) and -or(r) (-ør[r]) is the standard equivalent of Old Norse -orr. The Ostrobothnian form is knarr / knärr, which is also equivalent to knorr (but derived from Old Swedish *knarr, as u-mutation was unusual in Old Swedish).⁵³ A Northern Norwegian knorr is shown in Figure . The backward-curved stem is pronounced. It therefore makes good sense to deduce from the set of terms discussed above that knorr has the basic meaning of ‘steep or backward-curved’. This understanding seems to be confirmed by a comparison with the boat type prevalent in Trøndelag, the neighbouring area to the south of where the knorr was the traditional boat type. Until the early nineteenth century, the Trøndelag boat probably was the geitbåt, which is shown in Figure . It has a prominently projecting stem. The term geitbåt literally means ‘goat boat’ and is most easily explained as a comparison with a goat, which has a prominently projecting front part (a long neck). When viewed in light of the contrasting stems, both terms seem very characterising. In Ostrobothnia, we find a similar circumstance. A knarr / knärr / knärrbåt is shown in Figure  and has a stem that is quite similar to (albeit less pronounced ⁵³ It seems that eighteenth-century Shetland Norn also had the term knorr, referring to a small boat in local tradition, but we have no information about what characterised this boat (knorin definite form; Low : , cf.; Thowsen :  and Rendboe : –). 104 E H than) that of the northern Norwegian knorr. The other local boat type is the jullbåt, seen in Figure . The same stem contrast as on the western Norwegian coast is apparent. The fact that we find these contrasts in geographically distant areas strengthens the argument. The geographical separation means that the parallel can hardly be explained as a late borrowing and therefore rather derives from a common, medieval origin. Thirdly, the modern Norwegian adjective knarrstemnd ‘with a backward-curved stem’, literally ‘with a stem like a ‘knarr’’, has to be explained from the Old Norse linguistic stage because in Norwegian tradition there is no boat type *knarr (and there never was). Had the word knarrstemnd been formed from knorr in late tradition, it would have been *knorrstemnd, but that form is unknown. The word-form knarrstemnd, with an -a-, can only be explained from the Old Norse stem form knarr-, which was used in compounds (and retained the original a because in the stem form it was not followed by a mutation-causing u. The modern Scandinavian form knarr is a late borrowing from such compounds found in names of natural harbours where knerrir would lie: Knarrvik[a], Knarrlagsundet etc.). This means that the term knarrstemnd originates from Figure  (top): Eighteenth century knorr from Northern Norway. (Klepp : ) Figur  (øverst): -tals knorr frå Nord-Noreg. (Klepp : ) Figure  (above): Nineteenth or twentieth-century geitbåt ‘goat boat’ from Nordmøre, Norway. (Færøyvik : ) Figur  (over): - eller tals geitbåt frå Nordmøre. (Færøyvik : ) T  V   Figure  (top): Nineteenthcentury knarr / knärr from Swedish Ostrobothnia, Finland. (Ahlbäck and Bonns : ) Figur  (øverst): -tals knarr / knärr frå svensk Österbotten, Finland. (Ahlbäck and Bonns : ) 105 an unattested Old Norse form – *knarrstefndr – and consequently indicates that a backward-curved stem was considered a distinctive feature of a knorr in the Middle Ages. On the basis of the above, it is very probable that knorr originally meant ‘a ship with a backward-curved stem’. This characteristic fits well with the ships to the left in §  (Figure , Figure , Figure , Figure . Note the close resemblance between Figure  and Figure .), so it can be assumed that they were knerrir. This understanding of the term knorr may also explain how its meaning could shift from ‘warship’ to ‘cargo ship’. This shift is not a problem if knorr originally referred to the profile of the ship’s stem. In that case it did not originally express anything about cargo capacity, but a backward-curved stem could have become a characteristic of the big, ocean-going cargo ship when it emerged, because it is easier to build a broad, chubby hull if the stem is steep or backward-curved.⁵⁴ Skuldelev  (Figure ) from – AD is a fully developed ship of this type. It is widely considered to be a knorr (probably first suggested by Crumlin-Pedersen).⁵⁵ The reconstruction ⁵⁴ I have consulted several boat builders in the Norwegian west coast tradition on this. ⁵⁵ Crumlin-Pedersen : . 106 E H Figure : Probable knorr I-s. Left: warship on a Viking Age-pictorial stone from Gotland. (Hejnum Riddare, Lindqvist – no. ). Below: the Oseberg ship, Vestfold, Norway,  AD (poster published by the Museum of Cultural History, Oslo. Drawing by Lundin, ). Figur : Sannsynlege skip av typen knorr . Til venstre: krigsskip på vikingtids biletstein frå Gotland (Hejnum Riddare, Lindqvist – nr. ). Under: Osebergskipet, frå  (plakat publisert av Kulturhistorisk museum, Oslo. Teikning av Lundin, ) of the stem profiles shown in Figure  seems plausible,⁵⁶ and if so, it is *knarrstefnt. But it should be stressed that, in all probability, the term knorr only came to refer to cargo ships such as this as a secondary meaning. We should distinguish between ‘knorr I’, which would be ships like those in Figure  and those to the left in § ; and ‘knorr II’, which would be the later cargo ships like Skuldelev I (Figure ). ⁵⁶ …except that I doubt the basis for making the stem tops so short on the reconstruction. T  V   Figure : Probable knorr II. Cargo ship from – A.D, found in the Roskilde fjord. (Skuldelev . CrumlinPedersen and Olsen : ) Figur : Sannsynleg skip av typen knorr : Skuldelev , frakteskip frå –. (Skuldelev . Crumlin-Pedersen and Olsen : ) 107 The understanding of knorr as ‘cargo ship’ probably became an alternative meaning at some point during the tenth century, before eventually becoming the dominant meaning. This chronology seems to parallel that of Old English cēol. The meaning of this term also changed from ‘warship’ to ‘cargo ship’, the first certain example of the ‘cargo ship’ meaning dates from around  AD and it is therefore likely that it developed in the tenth century (see § ). This analogue supports the evidence of the two meanings of knorr, as well as the impression that specialised, ocean-going cargo ships in Northern Europe developed during the tenth century, whether they were referred to as knerrir, cēolas, or with other terms. The ‘cargo ship’ meaning of knorr never eradicated the other meanings. As we have seen, knorr survives as a term for traditional knarrstemnd vessels in Ostrobothnia and Northern Norway, whereas the ‘warship’ meaning of the word is still found in Faroese ballads, dating from the High Middle Ages or later (knørrur, knørur).⁵⁷ It is hard to tell when ‘cargo ship’ became the dominant ⁵⁷ Hammershaimb  II: , I e.g. , . 108 E H meaning, but it seems that an understanding of knorr as ‘warship’ survived for quite some time in skaldic poetry. The skald Steinn Herdísarson, probably around  AD, mentions steindir knerrir ‘painted knerrir’ among the gifts King Óláfr kyrri gave his followers.⁵⁸ As paint was very expensive, it is unlikely that cargo ships would have been painted, so these knerrir probably were warships – which also makes better sense in the context. Of the six eleventh-century Swedish runic inscriptions to mention knorr (above), none clearly refer to cargo ships and the majority refer to warlike situations.⁵⁹ In the High Medieval Swedish manuscripts, the term is not attested. This suggests that knorr never developed the meaning ‘cargo ship’ in Swedish (nevertheless, see the discussion in Jesch a: .). The two meanings of knorr complicate the understanding of the byname knarrarbringa. It is usually understood as a comparison between a ship with a broad fore, as in Figure , and a woman with a large, heavy, round bosom. This is unproblematic with regards to Ásný knarrarbringa, who lived in the High Middle Ages.⁶⁰ However, the above-mentioned Þorbjorg knarrarbringa allegedly lived in the ninth century and at that time it is very improbable that knorr referred to this kind of ship. Diverse and robust evidence indicates that a knorr in the ninth century was of the Oseberg type, which has a narrow fore. It is conceivable that knarrarbringa could also be derived from the appearance of this type, if it compares the posture of a woman bending backwards to compensate for her really heavy bosom with the profile of a knorr I, but this seems rather unlikely. It is therefore tempting to doubt the authenticity of the byname knarrarbringa in the ninth century. It could be a later construction. If there was a tradition that Þorbjorg was exceptionally well endowed, the ⁵⁸ Finnur Jónsson – B I: , ; Jesch a: . ⁵⁹ Varenius : –. ⁶⁰ See Jesch b. T  V   109 comparison of her father’s nose to a skeið could have motivated her own later comparison with a cargo ship or knorr II once it developed, much like Ásný and presumably other women not mentioned in the sources. This explanation of Þorbjorg’s knarrarbringa byname may look like ‘data massage’, but it is less problematic to doubt the authenticity of the byname than to argue for the ‘cargo-ship’ meaning of knorr already in the ninth century, because that would contradict many independent pieces of evidence. It is possible that the term knorr arose even before the Viking Age because we know of a distinct contrast in stem profiles before that time and because the other words of the type that knorr belongs to (e.g. fjorðr, skjoldr) are generally inherited at least from proto-Scandinavian (cf. the vowel shift above). The Norwegian Kvalsund ship from around  AD and the Migration Period rowing ships on pictorial stones from Gotland and Häggeby, Uppland, Sweden could thus be knerrir whereas the early seventh-century Sutton Hoo ship from England and the Nydam Ship, c.  AD, from southern Jutland would be the geitbåt type.⁶¹ These stem contrasts would actually be far more prominent than those we can observe in the Viking Age and it would thus make better sense if the term knorr were pre-Viking Age. However, our material does not allow any conclusions on the earliest days of the term knorr. . Beit Beit f. (pl. beit) does not occur in skaldic poetry, in bynames or in English or German sources from the early Viking Age. The term is mentioned nine times in the Old Norse poetic corpus, but none of these occurrences can reasonably be dated to before  AD.⁶² Still, I believe beit refers to an early Viking Age ship type, ⁶¹ Lindqvist – no.  etc.; Åkerlund : ; Nylén and Lamm : ; Engelhardt ; Rieck ; Gøthche . ⁶² Finnur Jónsson –: , cf. §  here. 110 E H because it never occurs in prose (see § ) and because it seems to be the only term that makes sense with a certain, very characteristic early ship type – namely the one with angular stems to the right in §  (see Figure , Figure , Figure ). In Heide and Planke, manuscript , I suggest this type was the beit, for three principal reasons: - The extensions under the stems appear simultaneously with the first sails, probably because sailing requires a better capacity to avoid leeway (when sailing upwind), and adding such triangles is the easiest way to achieve this on a rowing ship. This has been pointed out by Crumlin Pedersen, and the argument can be supported with Dutch tradition where exactly this kind of ‘skeg’ (in the fore of the ship; its potential place in the aft is occupied by the rudder because side rudders are abandoned) serves exactly this purpose.⁶³ - In modern Norwegian, beit refers to a ship’s ability to avoid leeway and this is probably related to the Old Norse verb beita ‘to sail upwind’, which seems to derive from an idea of the keel cutting through the sea, which in turn would fit well with this type of ship type and its extensions.⁶⁴ - The Dutch ‘skeg’ is called a loefbijter, literally ‘windward biter’, and the verb beita is the transitive form of bíta ‘to bite’. The essential meaning of bíta is ‘to cut’ (and beita, which is causative to bíta, means ‘to make / cause to bite’). Thus, the term loefbijter seems to support the connection between the term beit, the triangular extensions and sailing upwind. The argument for the identification of beit is far weaker than that given for knorr above or skeið below, but it ⁶³ Crumlin Pedersen (b: –; see Kruyskamp : . Thanks to Gerbrand Moyes for information on the Dutch loefbijter. ⁶⁴ For details, see Heide and Planke, manuscript . T  V   111 seems that this is the best option. No example of a ship with angular stems has yet been found, but due to the many depictions of them there is broad agreement that such ships really existed – some very detailed, as we have seen.⁶⁵ If this ship type existed, there must have been a term for it and, as it seems to have existed by the early Viking Age, it would be surprising if this term was not preserved in poetry. Of the candidate terms in the poetry, beit not only seems to be the only fitting alternative; it also makes good sense. The reason why beit is not mentioned in the earliest sources may be that it was a less important ship type and that this type predates virtually all of the preserved skaldic poetry. Our depictions of it are from the Merovingian period and the very earliest part of the Viking Age (the ship type seen on eleventh-century Swedish and Danish pictorial stones such as Ledberg and Törnevalla seems to be something else). Therefore, it fits well if beit only exists in our texts as a poetic term surviving from earlier times. It is conceivable that the very early attested barði (§ ) was an alternative term for this type of ship: apparently, it literally means ‘a ship characterised by its barð’ and the barð, defined as ‘the transition piece between the stem and the keel’, comes close to the part of the ship in question. However, it is uncertain whether barði is an early term because there is only one probable early attestation of it. . Skeið We have two indications that the term skeið f. (pl. skeiðar and skeiðr) existed in the ninth century. It is reflected in the byname skeiðarnef ‘skeið nose’, which was given to Þorbjorg knarrarbringa’s father Gils, who as earlier mentioned was probably born before  AD (§ .). If authentic, his daughter’s byname indicates ⁶⁵ See Heide and Planke, manuscript . 112 E H that skeiðarnef really is a comparison with the ship type. But there is reason to doubt the authenticity of Þorbjorg’s byname knarrarbringa (§ .), and therefore the skeiðarnef byname could originally have been a comparison with a weaving beater, which was also called a skeið f. (see Figure ). Even so, a comparison with a ship seems the more plausible alternative and skeið is also mentioned in the Danish runic inscription from Tryggevælde, dated to c. ; the term, accordingly, was established before that.⁶⁶ In the inscription, skeið is believed to refer to a ship setting, but as no ship setting is known in this place, it might be taken to refer to a racecourse, also called a skeið, instead. However, in the racecourse-meaning, skeið was neuter, whereas it was feminine when referring to a ship and in the inscription there is a congruence between skeið and a determinative that cannot be neutral: skaiþ þaisi (accusative). The whole sentence reads, normalised: ‘Ragnhildr, systir Ulfs, setti stein þenna ok gerði haug þenna ept, ok skeið þessa (skaiþ þaisi), Gunnulf, ver sinn, glomulan mann, son Nerfis…’, ‘Ragnhildr, Ulv’s sister, placed this stone and made this mound – and this ship setting (skaiþ) – in memory of Gunnulfr, her husband…’.⁶⁷ Accordingly, by far the most probable explanation would therefore be that skeið on the Tryggevælde inscription refers to a (stone) ship (which is now lost) and accordingly that a ship type skeið was established before c.  AD. The impression that skeið is an early type is supported by the fact that it is attested in tenth-century skaldic poetry (§ .) and is rare in the (prose of the) contemporary sagas written in the thirteenth century.⁶⁸ Crumlin-Pedersen suggests that the . m long Skuldelev  (built in , Figure ) and the perhaps  m long Hedeby  (built in ) are skeiðar.⁶⁹ This seems very likely (supported also by Simek ), and I ⁶⁶ ⁶⁷ ⁶⁸ ⁶⁹ Nielsen : ; Crumlin-Pedersen (: , a: ) Moltke : , my normalisation. Bjørgo : ; Malmros : . Crumlin-Pedersen : ; a: . T  V   Figure : Probable but late skeið: Skuldelev ,  AD. (Crumlin-Pedersen and Olsen : -) Figur : Sannsynleg, men sein skeið: Skuldelev , . (Crumlin-Pedersen and Olsen : -) 113 would like to add the  m long Roskilde , built after  and found in .⁷⁰ But Crumlin-Pedersen hardly gives arguments for the identification of skeið with these ships, although such arguments can be made. Firstly, a few skaldic poems, mostly from the late Viking Age, contain information that skeiðar are long and slender and this would fit with the Skuldelev  type, as pointed out by Jesch: they are langar ‘long’, súðlangar ‘with long strakes’, mævar ‘slender’, and þunnar ‘thin’ ( AD).⁷¹ Secondly, the byname skeiðarnef can most easily be interpreted as referring to a very long nose, which implies that skeiðar were long, too (but see below). Thirdly, from England, where the term in question was borrowed from Scandinavian, we have a mention of a scegð (pronounced /skeið/) with  oars, in bishop Ælfwold’s will from – AD.⁷² That makes  pairs – two more than on the reconstruction of Skuldelev , the Sea Stallion, which probably should have fewer oars given how difficult it has proved to row it fully manned.⁷³ Accordingly, a scegð /skeið could be a very long ship. Fourthly, the only plausible etymology of skeið is that it is a comparison with something long and narrow. Falk suggested that skeið is borrowed from Byzantine Greek ⁷⁰ Bill et. al.  and : ; Gøthche . ⁷¹ Jesch b: ; Jesch a: ; Finnur Jónsson – B II: ; Bjørgo : . ⁷² Thier : . ⁷³ Information from Tora Heide and others from the crew, September . 114 E H skhedía ‘a light ship’, but this is rejected by de Vries and Ásgeir Blöndal Magnússon because the word is attested in Scandinavian too early.⁷⁴ It has also been suggested that skeið means ‘a wave-splitting ship’ or ‘a ship made of planks made from split logs’, in both cases derived from the Germanic verb *skaiðan, ‘to split’.⁷⁵ However, this is not convincing (as Ásgeir Blöndal Magnússon points out) because all ships’ planks were made from split logs (before the saw), and all ships split the waves. Instead, Ásgeir Blöndal Magnússon suggests that skeið as a term for a ship type is a ‘re-use’ of a term skeið that referred to something long and narrow because the ship type in question had a similar shape. This seems plausible because the most striking feature of the ship type exemplified in Figure  is its extreme length and slenderness and we should expect that this would be reflected in a term designating this characteristic.⁷⁶ Ásgeir Blöndal Magnússon’s suggestion is that skeið as a term for a ship type is a comparison with skeið in the meaning ‘(sword) sheath’.⁷⁷ This is certainly possible, but perhaps even more plausible would be skeið in the meaning of ‘weaving beater’ (see Figure ).⁷⁸ I conclude ⁷⁴ Falk : ; de Vries : ; Ásgeir Blöndal Magnússon : . The explanation for the likeness may be borrowing in the opposite direction. ⁷⁵ See de Vries : . ⁷⁶ In principle, this term could also be dreki because the ships in question are long and slender like snakes, which came close to dragons in the Old Scandinavian mindset (footnote ). But we have no indication that the term dreki is older than the eleventh century and, as I argued above, it seems that dreki was an additional rather than a basic term. ⁷⁷ Foote is on to the same, : . ⁷⁸ Sheath is cognate with skeið. The connection to splitting is that a sheath was originally made by splitting a thin log, then hollowing-out the two pieces before tying them together. A weaving beater appears to ‘split’ the two layers of warp. Figure : Viking Age skeiðar ‘weaving beaters’.(Rygh and Lindberg , no. ) Figur : Skeiðar ‘vevsverd’ frå vikingtida. (Rygh and Lindberg , nr. ) T  V   115 that the Skuldelev  type, which we know from several ship-finds from the eleventh and late tenth centuries, most probably is the skeið. However, because the term skeið seems to have existed in the ninth century, the outlined understanding necessitates that long and slender ships also existed at that time. At first glance, this seems not to have been the case. Ole Crumlin Pedersen has pointed out that early Viking Age warships such as Tune and Gokstad are much broader compared to length than the eleventhcentury warships that we know of, and he has suggested that this is due to the need to ‘provide sufficient stability at the initial stage’ of sailing.⁷⁹ Only in the late Viking Age did warships return to the pre-Viking narrowness, Crumlin-Pedersen says. If this is correct, the term skeið could in theory derive from the long and slender rowing ships of the pre-sail era (like the -meter Sutton Hoo [Evans and Bruce-Mitford ] and the ships that were presumably stored in the well over  m long pre-Viking boat houses.⁸⁰ But this is speculation and not necessary, because our material does not permit the conclusion that no early Viking Age warships were slender. Unfortunately, the graffiti that show ships from the side give no information of how wide they were, but the Danish Ladby ship from – AD was , m long and had a length-to-breadth ratio of ,, which is quite close to Skuldelev ’s , and which is contrasted by Gokstad’s , and Oseberg’s ,.⁸¹ Thus, Ladby not only fits the understanding of skeið outlined above, but is also early Viking Age (Ladby is now dated – years earlier than when Crumlin-Pedersen published his article). To be sure, no such slender ship from the ninth century has been found, but this may very well be a coincidence as only two ships – Gokstad and Oseberg – definitely date from that century. ⁷⁹ Crumlin Pedersen a: –. ⁸⁰ Myhre ; Grimm ; Stylegar and Grimm . ⁸¹ Based upon Crumlin-Pedersen and Olsen : ; Sørensen : ,  ff.. 116 E H Even if slender warships did exist in the early Viking Age, there may be something to Crumlin-Pedersen’s idea. These ships were still a lot shorter than the longest rowing age ships, and really long ships did not return until the late Viking Age. Because the forces inflicted upon the hull by (the speed of ) sailing is far greater than the force produced by rowing, the longest rowing ships would hardly have withstood full-fledged sailing without fundamental changes to how they were constructed. But such changes cannot be made overnight, so the easy solution would be to make the longest ships shorter.⁸² To judge from the ship finds, it seems that designs for a sufficient longitudinal stiffness for sailing ships of great length and slenderness only emerges with the probable skeiðar from the late tenth and early eleventh centuries.⁸³ This development may be part of the reason why the skeið seems to become more important in the late Viking Age. It is not mentioned in Old English sources until then and it does not dominate the skaldic corpus’ references to warships until then.⁸⁴ It is also probable that (comparatively) broad warships went out of use in the late Viking Age (even if they were not the only type in the early Viking Age). This would fit in well with the specialisation of ship types that happened throughout the Viking Age. In the ninth century there were probably no specialised cargo ships, defined as ships with a large cargo capacity, virtually no rowing ability and just the crew to handle the ship.⁸⁵ The reason for this was probably partly that such ships had not yet had the time to evolve from the slender rowing ships after the introduction of the sail only one or two centuries before, and partly also that the Northern European states in the early Viking Age ⁸² Whether this really happened or not would probably become clear from the many Western Scandinavian boathouses, but, regrettably, too few of them are excavated and dated. ⁸³ See e.g. Bill, et. al., : . ⁸⁴ Thier : ; compare § . with Jesch a:  ff. ⁸⁵ First pointed out by Christensen . T  V   117 did not yet have a monopoly on violence, as CrumlinPedersen points out.⁸⁶ This would mean that a large, armed crew was needed in any case to protect the cargo and, consequently, that ships capable of making use of the rowing power of a large crew would be preferred. In the late Viking Age this was different and this development could lead to the abandonment of combined ships like Gokstad – warships with a large cargo capacity (in the High Middle Ages, this specialisation was reversed in some respects and new, heavy ship types used for both trade and warfare emerged). Here, I have adopted my predecessors’ view that (length and) slenderness were the main characteristics of skeiðar. From this, we cannot tell which of the ships in §  are skeiðar because these characteristics cannot be seen from the depictions. However, in light of the arguments presented in § , . and ., there is reason to believe that the skeið also had a distinctive stem profile. At least it is probable that the skeið had a different stem profile from the knorr, because if not, the term knorr would probably not make sense since it requires a contrast. If the term skeið really is a ‘re-use’ of skeið in the meaning ‘weaving beater’ or ‘sword sheath’, this could imply that the ship type skeið originally had a projecting stem, because weaving beaters and sword sheaths are pointed, as was in all likelihood the long nose of Gils skeiðarnef. If so, the vessel in Figure  could be an early skeið, as it has the most projecting stem of the known depictions from the early Viking Age. Its low freeboard would fit the slenderness of a skeið, because slenderness is normally connected to a low freeboard. However, our material is too limited to tell us whether the skeið originally had a projecting stem. ⁸⁶ Crumlin-Pedersen a: . 118 . E H Kjóll The term kjóll m. (pl. kjólar) is well documented in the West Germanic languages in the early Viking Age and even earlier. From England, cēol (pl. cēolas) is mentioned by Gildas in a Latin text as early as in the sixth century, referring to the long ships (longis nauibus) that first brought Saxons to England in the fifth century.⁸⁷ The later Anglo-Saxon Chronicle (ninth century and later) also refers to these first ships as cēolas or long ships and dates the event to the year .⁸⁸ In eighth-century poetry – for example Beowulf and The Fates of the Apostles – cēol is common, as it is throughout the Viking Age. In Old High German and Old Saxon, kiol appears in glossaries many times in the Merovingian and Viking Ages. In Old Norse, kjóll is mentioned as early as in Þorbjorn hornklofi’s Haraldskvæði (st. ) from the s or s (§ .) and in Þjóðolfr ór Hvini’s Ynglingatal (st. ) as well as a lausavísa by Þórir snepill, both probably from around  AD (§ .). Altogether, kjóll is mentioned some fifteen times in Old Norse poetry.⁸⁹ Norwegian farm names and the Old Norse form kjóll’s final, geminate -ll indicates that it existed in ProtoScandinavian and thus that *keul(az) was a common North-West Germanic term in the Migration Period or even earlier.⁹⁰ Kjóll is not found in Old Norse prose, however, with two exceptions that clearly or probably refer to English High Medieval (cargo) ships.⁹¹ Kjóll therefore seems to support the assumption that terms only found in poetry are survivals from the everyday language of an earlier period (cf. § ) – only that this particular term seems to have been re-connected in the High Middle Ages with the English cognate, which ⁸⁷ ⁸⁸ ⁸⁹ ⁹⁰ ⁹¹ Williams , ch. . Gildas’ spelling is cyulis, with a Latin dative plural. McCusker : ; Ellmers : . Finnur Jónsson –: . Heide manuscript . Falk : ; Olsen . T  V   Figure : Possible kjóll: the Gokstad ship, built c.  AD. (From Nicolaysen ) Figur : Mogeleg kjóll: Gokstadskipet, bygd ikr. . (Frå Nicolaysen ) 119 survived in English as a prosaic term for an English ship type at the time.⁹² It is commonly thought that the term kjóll / cēol / kiol is the same word as ‘keel’, ‘the bottom beam of a ship’, Old Norse kjolr m., but this is not correct; these words are not related, even if they merged in Middle English and the Middle German languages.⁹³ The cēolas of the Saxon settlers obviously were all-round ships – used for military attack, but fitted with a cargo capacity substantial enough to transport whole tribes across the North Sea. In the English poetry believed to date from the eighth century, cēol always refers to ships used for warfare.⁹⁴ However, dating this poetry is not easier than that of Eddic poetry, and since most of the poetic contexts in which cēol is used refer to warfare, this usage does not necessarily tell us much about what the cēolas were used for. But a connection between cēol and warfare in the early Anglo-Saxon era seems to emerge from Anglo-Saxon personal names like Cēolbald ‘cēol warrior’, Cēolhere ‘cēol ruler’, Cēolward ‘cēol guard’ etc.⁹⁵ However, from the late tenth century, possibly some⁹² ⁹³ ⁹⁴ ⁹⁵ Thier : . Heide manuscript . McCusker : ; Ellmers : ; cf. Bosworth and Toller : . Ellmers : . 120 E H what earlier, cēol / kēle started to refer to cargo ships.⁹⁶ On the continent it also seems that this word referred to warships in the Merovingian and Viking ages. In Old High German and Old Saxon, kiol appears in glossaries, translating Latin trieris ‘a galley with three banks of oars (on each side), trireme’, durco ‘big vessel’, and nauis magna ‘the same’, and celox ‘a swift vessel suitable for war service’. It also occurs in the meaning ratis ‘a raft, boat’, classis ‘a fleet’, and once, in the tenth century, in the meaning liburna, which in the Middle Ages referred to cargo ships.⁹⁷ According to Ellmers, kiol most often translates trieris, especially in the earliest period, the eighth and ninth centuries (the Old High German period goes to the mid-eleventh century). Because Gildas mentions cēolas in the sixth century, Ellmers and Thier connect the term cēol with the Germanic ships that we know from the age of the Germanic conquest of and immigration to England, this applies to the Nydam ship from around  AD, the Sutton Hoo ship from the early seventh century and the Snape and Ashby Dell ships.⁹⁸ This seems plausible, and we should probably add the Migration Period rowing ships of the Gotland pictorial stones and the Kvalsund ship from Western Norway.⁹⁹ There is agreement that an original form *keul-az with near certainty can be reconstructed on the basis of cēol / kiol / kjóll, but in Heide manuscript  I reject the idea that the original meaning was ‘a broad and high ship’, ‘a ship compared with a pot’, or similar, because such ships were not possessed by North⁹⁶ Ellmers (: –) argues that cēol could refer to cargo ships as early as in the eighth century, but the argument is not convincing. As pointed out by Thier (: ), the example from Andreas rather refers to a warship. The wine barrel riddle from the Exeter Book is hard to date. The earliest certain evidence of a cargo cēol seems to be the information from around the year  about the toll rate that cēolas were to pay (Falk : ; Thier : ). ⁹⁷ Ellmers : ; Karg-Gasterstädt and Frings -V: ; Gallé : ; Holthausen : . ⁹⁸ Ellmers : ; Thier : ; Engelhardt ; Rieck ; Gøthche ; Evans and Bruce-Mitford . ⁹⁹ Lindqvist – no.  etc.; Åkerlund : ; Nylén and Lamm : ; Shetelig and Johannessen . T  V   121 ern Germanic peoples until centuries later. Instead, I suggest a more general meaning of ‘container’, as the closest related terms are Old Norse kýll ‘a sack, bag’, and Old High German and Old Saxon kiula ’a bag’, both from *keul-. This meaning would make sense because the ships that brought the Saxon settlers across the North Sea were obviously the best cargo carriers that they had, even if by later standards they carried little because rowing ships must be low and slender. In that case, *keulaz could be a generic term covering several designs of large, long-distance rowing ships, as the early Germanic ships include both the knorr and the geitbåt stem profiles (§ .), and it would be surprising if they were not designated with additional terms indicating this. However, the focus of discussion here is the kjólar of the early Viking Age and as specialised sailing ships they must have been quite different from the rowing ships of the Migration Period. Accordingly, we have to study the information that we have on the early Viking Age kjólar. It is quite limited, but may still be sufficient: according to Haraldskvæði  (§ .), the kjólar used in the battle of Hafrsfjord in the s were djúpir ‘deep’. This indicates that a comparatively high freeboard and so a comparatively large cargo capacity was a characteristic of kjólar. Stanza  in the Eddic poem Voluspá, generally assumed to have been composed around the year ,¹⁰⁰ points in the same direction, because it says that a kjóll carries all the world’s jotnar to the Ragnarok battle (alliteration with koma in the next line supports the assumption that kjóll really belonged to the early version of the poem). It also corresponds to the fact that the English cēol developed into a cargo ship (§ .), because this development is easiest understood if its startingpoint was an all-round ship with a large cargo capacity (although the case of knorr reminds us that the continuity behind the term could be represented by another trait than cargo capacity). In fact, the cēol seems to have ¹⁰⁰ E.g. Simek and Hermann Pálsson : . 122 E H been conceived of as the cargo carrier in pre-Viking times (even if it must then have been low and slender compared to later times), judging by the etymology and the fact that it was chosen for the conquest and settlement of England across the North Sea. Taking everything into account, there are quite a number of indications that seem to point to an early Viking Age kjóll having a comparatively high freeboard and a comparatively large cargo capacity. If this is correct, the best candidate for identification is the Gokstad ship, built around  AD. Gokstad has a hull of the type in question and its stem profile is ‘vacant’, cf. § . and .. The depictions in Figure  and Figure  may also be of this type. The results of this discussion indicate that the kjóll was the ship type that primarily brought the settlers to Iceland with their livestock in the late ninth century and the early tenth century, as the kjóll seems to have been the ship most suitable for overseas cargo transport in the period before specialised cargo carrier knerrir, like Skuldelev  in Figure , was developed – but see § . In Scandinavia, the term kjóll fell out of use. It is hard to know when this happened, but it is natural to relate it to the development of the cargo carrier knorr. In that case, the term went out of use because its referent was replaced by a new ship type, the knorr II, which was broader and higher – and was knarrstemnd (§ .). This may have happened between the mid-tenth and the mid-eleventh centuries. . Askr The term askr m. (pl. askar) is only attested twice in Old Norse poetry, in skaldic poems dated to the late tenth century (§ .), which is quite early.¹⁰¹ In prose there is a handful of occurrences in two texts, Hervarar saga ok Heiðreks (, ch. ) and Örvar-Odds saga (, ch. ), ¹⁰¹ Finnur Jónsson –: –. T  V   123 which are accounts of the remote past. This indicates that the askr was not a major type, but still early, and that the term was no longer in use in everyday speech in the High Middle Ages. At first glance, the prose occurrences seem to indicate the opposite, but as askr is never found in the contemporary sagas the instances where it is used are most easily understood as deliberate archaisms (cf. § ), like the poetic term sætré ‘sea tree’ / ‘a ship’ used in prose in one of these sagas, Örvar-Odds saga (ch. ). Seemingly, the authors of these sagas understood askr as an ancient type of ship. This interpretation is supported by West Germanic sources where cognates of askr are recorded before and during the (early) Viking Age.¹⁰² The sixth-century Frankish law Lex Salica (earliest manuscript c.  AD) mentions the Latinised form ascus and states a substantially higher fine for the theft of a locked-up ascus that was hanging up for inspection than for an ordinary boat (nauis). Accordingly, an *asc was valuable and light.¹⁰³ Kuhn understands the *asc as a river vessel – its lightness would fit with that, the Salian Franks were a continental people living around the upper Rhine and the upper Danube (present-day Bavaria and Baden-Württemberg), and when the *asc reappears in sources in the Late Middle Ages (as asch) it is a river boat in this area. In Old English, æsc is attested in all of the three early glossaries (Corpus, Épinal, and Erfurt) and therefore probably existed already in the assumed source for these glossaries, in the early seventh century.¹⁰⁴ Here, æsc translates cercylus, which was a light and swift vessel, thus similar to the Frankish *asc. Therefore, and because æscas are never mentioned in connection with the Migration Period invasion of England, Kuhn argues that the indigenous English pre-Viking æsc ¹⁰² It may also be supported by a certain Oddbjorn’s byname askasmiðr. He is mentioned in Landnámabók (: ) and Egils saga (: , footnote) among the first two generations in Iceland and the byname is most easily understood in the meaning ‘builder of askr ships’, as a parallel to knarrarsmiðr and skipasmiðr, but it could also mean ‘spear-maker’ or ‘bentwood box maker’. ¹⁰³ Kuhn ; Thier : . ¹⁰⁴ Kuhn , Thier : . 124 E H was a river vessel, like the Frankish *asc.¹⁰⁵ In the Viking Age, however, æsc exclusively refers to Scandinavian warships in English sources. In this meaning, it is mentioned two times already in the ninth century ( and ), but only a few times altogether. Still, Kuhn argues that it must have been a major type in the early Viking Age because ascmann ‘a Scandinavian pirate’ (literally: ‘*asc man’),¹⁰⁶ parallel to Old English æscman ( AD), is also attested in the Low Countries and Northern Germany, in the late Viking Age.¹⁰⁷ If so, the reason for the few mentions of æsc / *asc / askr is that early Viking Age sources in the vernacular are few and that the oceangoing æsc / *asc / askr fell out of use after the early Viking Age (although the compound æscman / ascmann lived on in the secondary meaning of ‘Viking, Scandinavian’). All of Kuhn’s considerations seem plausible, but the askr still remains confusing. The etymological identity of askr, *asc, and æsc, and the early attestations of the West Germanic forms, as well as the geographical separation between them, all point to *askaz having been a common Northwest Germanic type in the Roman Iron Age and the Migration Period. This is not contradicted by the fact that the Scandinavian version was oceangoing, while the West Germanic versions seem to have been river vessels, because the Scandinavian version is recorded later and may have been more developed at that point. But in that case, why did the English still identify the Scandinavian ships with their own (probably) river-going æscas, and why did the æsc / askr seemingly go out of use after the early Viking Age? The commonplace etymology does not help us answer these questions. It claims that the term æsc / askr reflects the building material: this ship type was built of ash wood. But this is very unlikely because ash wood was only exceptionally used for shipbuilding (as many ¹⁰⁵ Kuhn . ¹⁰⁶ Kuhn . The cognate askmaðr is found as a byname in Old Norse (Fritzner – I: ), but it seems to give no information of the ship type askr. ¹⁰⁷ Kuhn . T  V   Figure : Norwegian sveipask / øskje with a lid, both sewn together with roots. (Photo: Arctandria, Tromsø) Figur : Sveipask / -øskje, sydd i hop med rottæger. (Foto: Arctandria, Tromsø) 125 scholars point out), because it is hard to split into material for planks.¹⁰⁸ But ash is very suitable for bentwood containers, which are made by bending a (green or steamed) thin board, to form the wall of the container and stitching it together (before putting a bottom inside of it), see Figure . This technique and the ash wood’s suitability for it is probably why a wooden container could be called ‘an ash’ from very early on in Germanic. The alternative – that ash wood was preferred for carving out bowls – is unlikely because ash wood is quite difficult to carve. In Norway, the connection between the term ask and the bentwood technique is still preserved.¹⁰⁹ The form and distribution and early attestations of several of the ‘ash’ terms for a wooden container indicate that *askaz in this meaning existed in Northwest Germanic: askr m. in Old Norse, asc (with variants) in the Middle German languages, eski n. in Old Norse and øskje in modern Norwegian (< *askijōn).¹¹⁰ The relationship between the meanings of *askaz as container and boat is usually explained as the former being derived from the latter, because a wooden container resembles a boat.¹¹¹ However, it is rather the other way around: the starting-point is the suitability of ash wood for bentwood boxes; the ship term *askaz is a ‘re-use’ of the term for a bentwood box¹¹² and it referred to ships with (some of ) the planks sewn together (like on a bentwood box) rather than joined with iron rivets. This stitching is a major characteristic of sewn boats and it is easily visible, even at a distance, as can be seen in ¹⁰⁸ Ásgeir Blöndal Magnússon : ; Kuhn (); Thier (: ; Osborn ; Sayers . ¹⁰⁹ Aasen : . ¹¹⁰ Fritzner – I: ; Kluge and Seebold : . ¹¹¹ E.g. Kluge and Seebold : . ¹¹² Sayers also suggests that the ‘container’ meaning is the starting-point, but he is unaware of the connection between askar and the bentwood type of wooden container, so he suggests that the ship type askr resembled a shallow bowl, which seems unfounded (Sayers : ). 126 E H Figure , so we should not be surprised if this characteristic gave rise to a term for a ship type. This explanation of askr may initially seem unlikely because it is commonly assumed that sewn ships went out of use before the Viking Age. But the picture is more complicated than that. It is true that already the Nydam ship from  AD was riveted and that all the famous Viking ships are riveted. Still, sewn boats and ships have remained in use in Northern Europe up to modern times.¹¹³ The technique of sewing survived longest on river boats because it is lighter than riveting, which is an advantage when passing portages, but sewn ocean-going cargo carriers are known as late as in the seventeenth and eighteenth centuries, from the Baltic and the Barents Seas.¹¹⁴ In the period from which we have evidence, the sewing technique was pushed back towards the east and north in Northern Europe, with long periods of transition, and sewing and riveting often combined on the same boat or on different types.¹¹⁵ ¹¹³ Westerdahl b, a. ¹¹⁴ Westerdahl b:  ff.; Westerdahl a: , ; Tanaka . ¹¹⁵ In the western Scandinavian languages, the clinching of a ship’s board is still called a súð / su ‘sewing’, and a boat rivet a saumr ‘seam’ (in Danish, søm has even assumed the general meaning ‘a nail’). Figure : Sewn boat: replica of a schnjaka from the Kola peninsula – . metres long original from c.  at the Norwegian Maritime Museum, Oslo. (From http://www.sewboat. narod.ru/shnjaka/) Figur : Kopi av schnjaka frå Kolahalvøya. Original frå ikr. . (Frå http://www. sewboat.narod.ru/shnjaka/) T  V   127 In Viking Age Scandinavia, we know of several (partly) sewn boats from the coast of Nordland and Troms counties, Northern Norway; the largest being the Bår(d)set boat from – AD, probably c. thirteen metres long and with seven pairs of oars.¹¹⁶ But sewn boats are also known from southern parts of Norway and Sweden at that time: eighth-ninth century Västmanland, central Sweden; mid-ninth century Sunnmøre, Western Norway (a sewn repair, Fjørtoft); and eleventh-century Trondheim, Central Norway.¹¹⁷ The term taghbænda mentioned in the provincial law of Västergötland, South-western Sweden, anchors the sewing technique to this area, too, as late as in the thirteenth century.¹¹⁸ An account in Heimskringla mentions two sewn ships with twelve pairs of oars built by the Sami in Nordland as late as in –. Some of the Norwegian archaeologically found sewn boats, especially those from Northern Norway, have been interpreted as Sami because they are sewn (see footnote ). This may be right, but they are found in Germanic contexts and, in the Viking Age, the sewing technique was not yet restricted to the Sami, but was used by the Germanic population even in Southern Sweden and Southern Norway. In  a Greenlandic ship ‘held together with wooden pegs and baleen or sinew lashing’ came to Iceland.¹¹⁹ There probably is reference to sewn ships even in the Old English poem Beowulf ¹²⁰(line –: guman ūt scufon, / weras on wilsīð wudu bundenne, ‘the fellows shoved off / men on a desired voyage, in a bound ship’, possibly line : bundenstefna, ‘bound prow’).¹²¹ This ¹¹⁶ Westerdahl b: , a:  ff.; Pedersen : , –. ¹¹⁷ Tuna, Badelunda, Westerdahl b: , a: –; Nordeide : . In the book, Nordeide mentions that according to Arne Emil Christensen, boats were sewn in Sami tradition, thus implying that this boat was Sami. But Christensen also said that it was divided into roms (Old Norse rúm) according to Norwegian rules (personal communication from Nordeide January ). ¹¹⁸ Äldre Västgötalagen : , cf. Steen : ; Westerdahl a: . ¹¹⁹ Ólafur Halldórsson : , , –. ¹²⁰ Thanks to Katrin Thier for pointing this out to me. ¹²¹ Westerdahl b, a; Fulk et. al. : , ; clxix, , , ; Thier : . 128 E H indicates that this technique was still known in eighthcentury England, but these phrases could also be taken to reflect an earlier technology, preserved in the conservative, poetic language. On the background of the above, it is plausible that some or many of the ships used by the Scandinavians in the first phase of the Viking raids were sewn. To be sure, no sewn (ocean-going) ship from the early Viking Age has been found in Southern Scandinavia, but this may be the result of accidental archaeological preservation. Our material is so limited that it reflects only a fraction of past reality and sewn ships must be underrepresented. In many Viking Age ship finds, virtually the only thing to survive is the rust from rivets. If these boats had been sewn it is unlikely that they would ever have been found, since the stitching would leave no more trace than the wooden planks of the boat it held together. In addition, it is likely that riveted ships were preferred as grave goods because iron was expensive and thus reflected a higher status. As all the ships so far found from the early Viking Age have been grave ships, this may have made a big difference. It is also worth bearing in mind that many of the ships that were used for warfare in the early Viking Age must have been smaller than Oseberg and Gokstad, which we tend to understand as ‘typical’ early Viking Age (war)ships. Both are so big that they have oar-holes rather than oarlocks with oar-grommets, but according to Þorbjorn hornklofi’s Haraldskvæði (st. ) oarlocks and oar-grommets were broken on ships during the battle of Hafrsfjord in the s (homlur at slíta / en hái at brjóta).¹²² This points to quite small ships, perhaps as small as the Bår(d)set boat, and the Scandinavian æscas / askar recorded in England were small (see below).¹²³ ¹²² Finnur Jónsson – B I: , cf. § .. ¹²³ According to de Vries (: ), Old Scandinavian askr was borrowed into Byzantine Greek askós ‘a ship’, but in the dictionaries, there is no trace of a ship askós, so it cannot help us (thanks to Gjert Vestrheim for this information). T  V   129 If the *askaz was a sewn boat or ship and the term derives from a comparison with bentwood boxes (which are sewn), the term in all likelihood emerged with the invention of riveted ships, to make it possible to distinguish between vessels built with the two joining techniques. If so, *askaz would be an additional term, referring to a ship’s building technique rather than its design, and an *askaz could at the same time be referred to with other terms that designated other characteristics, such as *keulaz, or, in the Viking Age: knorr, beit, or other (cf. § ). One point in favour of this theory that goes above and beyond the evidence discussed so far is the fact that it can provide answers to many questions. Firstly, if (Scandinavian) askar were ships with sewn planking, it could explain why the English probably identified them with an English type of river boat: this is not problematic if both types were characterised by the sewing technique although they were different in other respects. This understanding presupposes that, by the early Viking Age, no oceangoing craft were sewn in England, but that (some) river boats still were. This seems a fair assumption given that the technique was gradually being pushed back to the north and east, that sewn ships are probably mentioned in Beowulf, that the technique is attested in much of Scandinavia at this time, but alongside the riveting technique, which dominated; and that the sewing technique invariably survived longest on river craft that needed to be light because of portages. Secondly, this theory can explain why the Frankish *asc was so light that it was commonplace to hang it up for inspection (its high value suggests that it was still quite large and stately, perhaps resembling the late Roman warships that we know from the Rhine).¹²⁴ Thirdly, if the askar / æscas were sewn, this could explain the problematic term nægled cnear in the Anglo-Saxon Chronicle entry from shortly ¹²⁴ See Höckmann ; we know that Charlemagne kept a fleet on the Rhine and the Danube in the eighth and ninth centuries [Hausen ]. 130 E H after  (§ . here): after a battle, the Vikings ‘put out in their nægled cnears [nailed cnears] on to the sea’.¹²⁵ Why would the poet want to tell us that the Vikings’ cnears were nailed? The explanation may lie in what King Ælfred did some  years earlier ( AD): he gave instructions that a new type of warship was to be built ongen ða æscas ‘against the askar’.¹²⁶ These ships ‘were almost twice as long as the others, some had sixty oars, some more; they were both swifter, steadier, and with more freeboard than the others’. This implies that the æscas / askar were small, and the king’s action inevitably would trigger an early phase of the arms race that was to accelerate in the High Middle Ages, with an increasing focus on bigger and higher ships specialised for sea battles¹²⁷ rather than beach landings, which the Vikings could without any obstruction have given priority up to this point. The Vikings would have to respond and this would have forced them to give up the smallest ships, those most suitable for beach landings and consequently also for the sewing technique, which is only applicable to small ships. Therefore, it could be worth mentioning  years later that (now) the knerrir were nailed (and not sewn any more; they were no longer æscas). Finally, this would explain why the askar fell out of use after the early Viking Age. . Elliði Elliði m. (pl. *elliðar) is rare in the Old Norse corpus. In poetry, it is only attested in a lausavísa by Kormákr O ˛ gmundarson from –, in Eilífr Goðrúnarson’s Þórsdrápa  from around  AD (§ .) and in Eddic-metre poetry in the late Friðþjófs saga.¹²⁸ In prose it is attested in Landnámabók and in late sagas about the legendary past: Hversu Noregr byggðist, Friðþjófs saga, ¹²⁵ ¹²⁶ ¹²⁷ ¹²⁸ Battle of Brunanburh, Livingston : , , cf.; Whitelock et. al. : . Two Saxon chronicles : . See Sverris saga . Friðþjófs saga  ch. . T  V   131 and Sorla þáttr.¹²⁹ In Landnámabók, Friðþjófs saga and Sorla þáttr, Elliði is a proper noun, in the latter two texts referring to Friðþjófr’s magical ship.¹³⁰ Such meagre evidence could suggest that elliði should be ignored in this discussion because the type was marginal or did not belong to the early Viking Age. But there is additional material indicating that it was not all that marginal as well as many indications that it was early. With regards to the latter: firstly, the fact that elliði / Elliði is mentioned in several legendary sagas, even if it never occurs in the contemporary sagas or High Medieval poetry, indicates that the writers of the legendary sagas understood elliði as a ship type belonging to ancient times. Secondly, the High Medieval Icelandic tradition connects elliði / Elliði with the first two generations of Icelanders: according to Landnámabók, Elliði was a ship owned by Ketilbjorn gamli Ketilsson, who settled in South-western Iceland.¹³¹ Elliða-Grímr was the son of Ásgrímr O ˛ ndóttsson, who settled in Eyjafjörður.¹³² Álof elliðaskjoldr ‘elliði shield’ was the daughter of Ásgerðr Asksdóttir, who settled in the south-west of Iceland.¹³³ Thirdly, elliði occurs in quite a number of Icelandic place names, many of them mentioned in Old Norse texts, ¹³⁴ and as elliði seems not to have been in common use in the High Middle Ages, it is likely that these names were given before that. One of them, Elliðaárós ‘elliði river mouth’ in present-day Reykjavík, supports the tradition of Ketilbjorn gamli and his Elliði, because according to Landnámabók this is where he landed.¹³⁵ Elliðaey ‘elliði island’ in Breiðafjörður is also mentioned in connection with the settlement.¹³⁶ Combined, these ¹²⁹ Landnámabók : ; Hversu Noregr byggðist , ch. ; Friðþjófs saga , in numerous places; Sörla þáttr : . ¹³⁰ Friðþjófs saga , ch. . ¹³¹ Landnámabók : . ¹³² Landnámabók : , , , ; Njáls saga  in many places. ¹³³ Landnámabók : , , –. ¹³⁴ Svavar Sigmundsson . ¹³⁵ Landnámabók : . ¹³⁶ Landnámabók : . 132 E H arguments give quite strong reason to believe that the ship type elliði was in use when Iceland was settled. The place-names and bynames indicate that it cannot have been that unusual either. The question remains as to what kind of a ship the elliði was. Falk argues that the term elliði is borrowed from Slavic peoples and compares it to Old Church Slavonic alŭdija, ladija ‘a boat’ and Lithuanian eldija, aldija ‘a barge, a lighter’.¹³⁷ Old Norse leðja f. has the same origin and is a variant of elliði, according to Falk. In Old Norse, leðja is only attested in the Icelandic thirteenth-century Þulur list of terms for ship types, with no meaning given, but it is attested around the same time in Old Swedish, where lädhia / lodhia / lydhia / lydhugia has the meaning of ‘some kind of flat-bottomed boat used by the Russians’.¹³⁸ In Medieval Novgorod, a lodija was a ‘Russian river, lake and sea vessel’.¹³⁹ In Middle Low German, a lod(d)ie or lod(d)ige was a small cargo ship mainly used in transportation on the Neva and Volkhov rivers (up to Novgorod) and along the coast to Riga, or ‘a shallow vessel, a barge’.¹⁴⁰ In Late Medieval and Early Modern Swedish, lodja refers to ‘some kind of flat-bottomed rowing vessel used on shallow waters, particularly in Russia and Finland’.¹⁴¹ In Estonia, the lodi was an extremely broad, slow cargo carrier used on rivers and lakes into the early twentieth century (Past ). In Norwegian dialects lorje / lørje refers to ‘a (big), open, flat-bottomed boat, a barge’.¹⁴² Early Modern Danish lorje refers to ‘a small, flat-bottomed boat or barge in particular used as a cable ferry’.¹⁴³ Most scholars agree with Falk that the elliði and the leðja are borrowed from Slavic peoples and / or other ¹³⁷ Falk : . ¹³⁸ Footnote ; Finnur Jónsson – B I: ; Söderwall –: , , , . ¹³⁹ Sorokin : , cf. : . The spelling lodija from de Vries : . ¹⁴⁰ Cordes et. al. - II: ; Schiller and Lübben  II: . ¹⁴¹ Ordbok över svenska språket  - : . ¹⁴² Norsk Ordbok - . ¹⁴³ Ordbog over det danske sprog – : . T  V   133 peoples east of the Baltic Sea – for good reasons. The formal linguistic connection between leðja / loðja and Russian lodija seems very probable. The connection between elliði and alŭdija / eldija also seems plausible and there is no better explanation for elliði. It has been suggested that it derives from *ein-liði ‘solo sailor (because it is so fast)’,¹⁴⁴ but this seems rather farfetched and there is no indication that the elliði type was particularly fast. Leðja, however, can hardly be a variant of elliði. It is rather a later loan from the same source, because leðja / loðja does not seem to be older than the thirteenth century (in Scandinavia), while elliði appears to belong to the early Viking Age, as we have seen. Falk does not discuss why the Scandinavians would borrow a term for a ship type from the Slavs (and / or the Balts), but if we want to get closer to the elliði, this question is essential. At first glance, this borrowing is surprising because loanword studies teach us that if the Scandinavians borrowed a term for a ship type, they most likely also borrowed the ship type. During the period in question, however, the Scandinavians dominated Northern European maritime technology and it seems that only exceptionally did they borrow technical solutions and terms, while other peoples frequently borrowed from them. On closer inspection, however, we should expect that the (Northern) Slavs were the foremost experts on river vessels because they lived in vast expanses of land mostly connected via boat travel on enormous river systems. Accordingly, it should not surprise us if the Scandinavians borrowed lake and river boat types or combined river / sea vessel types from the Slavs, and of this we have examples: the type prámr (today pram) m., first attested in Old Norse in Þulur, was ‘a flat-bottomed rowing boat; a barge’, and the term is borrowed from Slavic.¹⁴⁵ It is likely that so, too, was the term karfi, which seems to have referred to a com¹⁴⁴ Alexander Jóhannesson :  in de Vries : . ¹⁴⁵ Finnur Jónsson – B I: ; Falk : . 134 E H bined inland / ocean-going vessel in the Scandinavian High Middle Ages .¹⁴⁶ The leðja and the elliði seem to be two more examples. As we have seen, the Russian lodija was a combined inland / sea vessel and the Medieval Swedish and later Scandinavian lodja / lorje was a flat-bottomed craft suitable for shallow or sheltered waters. There are also indications that the elliði was such a combined vessel, namely in the tradition connecting the elliði to the Namdalseidet isthmus between Trondheimsfjorden and Løgnin / Namsenfjorden to the north-west, in Central Norway. Ketilbjorn gamli is said to have come to Iceland with the ship Elliði from this area (Namdalen) and, according to Hversu Noregr byggðisk, the legendary King Beitir fixed an elliði to a sled in mid-winter and sailed across this isthmus because he had been promised all the land to the port side of his ship as he sailed north up the Norwegian coast.¹⁴⁷ This is unhistorical of course, but there is reason to believe that inherited oral tradition lies behind the tale, and why state that the king used the unusual ship type elliði in this operation? Was it because elliðar were especially suited for isthmus crossings?¹⁴⁸ The isthmus is not dry all across; for most of the journey one would let the vessel float in a small river with the telling name Ferja ‘the ferry’ – provided one had a relatively flat-bottomed vessel that was small enough to suit this kind of combined journey. Namdalseidet is by far the widest of the frequently used isthmuses on the Norwegian west coast – it is now close to twenty kilometres across (after one more millennium of post-glacial rebound) – and there is sea on either side, rather than a lake to one side, as is the case with the major isthmuses in Eastern Norway. Close to the isthmus on the north side is the mouth of one of Norway’s longest rivers, Namsen, Old Norse *Nauma ¹⁴⁶ As I will argue in a forthcoming article, cf. Falk :  and Korhonen . ¹⁴⁷ Flateyjarbok – I: , Hversu Noregr byggðist , ch. . ¹⁴⁸ According to Orkneyinga saga (: –), King Magnús Barefoot did the same trick on the west coast of Scotland in the latter part of the eleventh century, with a skúta, which was a small and light ship (common in the High Middle Ages) and thus also suited for this use. T  V   135 (the valley in which it flows was called Naumudalr ‘*Nauma valley’), which probably means ‘the boat river’ (from nau- as in Old Norse naust ‘boat house’ and Latin nauis ‘a boat’).¹⁴⁹ For these reasons, Namdalseidet is the place in Norway most likely to support a specialised ship type for the combination of sea, river, and isthmus dragging. This may be why Ketilbjorn gamli owned an elliði: it was more useful for him than for most others since he lived in an area that called for such a vessel, and therefore he sailed to Iceland in one (I find it more plausible that Ketilbjorn had an elliði than that he had a ship named Elliði; when elliði was a ship type, this term could hardly have functioned as a proper name without some kind of addition – rather like naming a ship *Tanker or *Cruise Ship today. Therefore, there is reason to believe that the elliði, traditionally connected to Ketilbjorn gamli, was reinterpreted as a proper name once this ship type fell out of use; this Elliði may have inspired the name of Friðþjófr’s ship). If the elliði was a combined river / sea vessel, this may also explain why it is so rare in Old Norse poetry: such vessels are often considered ugly and not proper boats or ships, whereas skaldic poetry shows great preference for high status topics. This understanding fits in well with the only two skaldic references to elliðar; they are both found in the context of mocking. The Elliði of Friðþjófs saga, on the other hand, is highly praised, both in the prose and the Eddic-metre poetry, but it is hard not to see this as a product of fantasy at a time when the memory of actual elliðar had faded. From this discussion, it seems that the elliði can with some probability be understood as a rather flatbottomed vessel that combined both river and sea use and that was not uncommon in the early Viking Age. Yet it seems impossible to give more detail and to identify Viking Age ship depictions or finds as elliðar. ¹⁴⁹ Sandnes and Helander . See Finnbogi Guðmundsson (: , footnote ), about the variant Naumsi, which does not necessarily contradict the essence of the above. 136 . E H Conclusion All the results presented in this article have some degree of uncertainty to them and, because of the source situation, it is unlikely that we will ever reach complete certainty with regard to these questions. Even so, the study is valuable, because the results on knorr and skeið are quite certain, and because what should be required from research is only that it presents understandings that are more probable than existing alternatives, by being based on as firm evidence as possible. Our view of the Early Viking Age ship types should be based upon collection and systematisation of all the available data, however insufficient it may be, not on guesswork and High Medieval misconceptions. What we can hope for to improve the source situation is more ship finds, especially from the Baltic Sea, which has hitherto not had shipworms, because of the brackish water, so gunwales, stems, and other term-defining traits may still be intact. Denmark took the lead in Viking ship research after the find of the Skuldelev ships in the s, but this position ought to be taken by the countries around the northern Baltic because they have the best natural conditions for the preservation of wrecks. To my knowledge, wrecks of Viking Age ships have already been identified on the Swedish coast, but lack of interest from funding bodies blocks examinations – and now a new type of shipworm is spreading rapidly, so time may be running out. Thanks to Judith Jesch, Arne Emil Christensen, Katrin Thier, Karen Bek-Pedersen, and the participants at the Bergen Old Norse Research Group’s seminar in November , for comments on drafts of this article. T  V   137 . Appendix. The earliest Scandinavian evidence of terms for ship types Compare the discussion in § . Although I define the middle of the tenth century as the end of the early Viking Age, I include attestations up to around  AD, to make sure that all relevant attestations are included and as a background for the discussion. Attestations from the second half of the tenth century have not been used as part of the basis for conclusions. The terms given in bold type are those with which the discussion in §  concluded. . Ship types and ship designations in Old Norse skaldic poetry until  AD The list was compiled by: - going through the normalised and corrected volume (B I) of Finnur Jónsson’s edition of the skaldic poetry (–), up to ‘The  century’ (p. .),¹⁵⁰ - consulting all the ship type terms mentioned by Falk () in Finnur Jónsson’s Lexicon poeticum (–) and checking these attestations in the edition (B I),¹⁵¹ - checking in the variant volume (A I) that the occurrences are valid. Kennings, such as drasill sunda ‘waterway steed’ and brimdýr ‘surf animal’, are not included. Naðr and Ormr ‘snake, dragon’ are not listed when they are short names for the ships Ormrinn langi and Ormrinn skammi. ¹⁵⁰ I use Finnur Jónsson’s edition because the on-going Skaldic Project had not reached the oldest poems when the compilation was done and does not present the skaldic corpus chronologically. ¹⁵¹ Falk :  ff. 138 E H FJ page 4 Bragi gamli Boddason FJ’s Name of poem dating Ragnarsdrápa 17 4 Bragi gamli 800 –850 7 11 Þjóðolfr ór Hvini -”- (Late?) Ynglingatal 4 9 c -”-”- 24 11 -”- -”- -”- 24 19 -”- -”- Lausavísa 2 20 Þorbjorn hornklofi c. 900 Glymdrápa 2 20 22 -”-”- 23 -”- -”870s or 880s? -“- 29 Þórir snepill 31 31 Egill SkallaGrímsson -”- 34 41 -”-”- 42 -”- 47 55 -”Guttormr sindri -”- 56 65 Glúmr Geirason ‘Ubestemmelige vers’ -”- 3 Haraldskvæði 5 Term / quotation The sea referred to as the road of a borðróinn barði ’a ship rowed on both sides’. Þars sem lofðar líta lung váfaðar Gungnis ’it was as if the men saw Óðinn’s lung’. Uncertain. We have no information that Óðinn had a special ship, so we should rather expect that lung here refers to his horse, Sleipnir. ‘House’ = Arinkjóll ’hearth ship’. ‘House’ = brandnór ’hearth ship’. ‘House’ = toptar nokkvi ’toft ship’. ‘Sea’ = flatvollr fleya ‘the flat field of ships’. ‘Fleet of warships’ = rosinaðr ‘splendid dragon’ and no˛ kkvar (pl.). ‘Warship’ = skip. ’Warships’ = kjólar Finnur dates Þorbjorn, (pl.). to ‘around 900’. See § 6. -”- 7 ’Warships’ = knerrir (pl.) c. 900 Lausavísa ‘Warrior’ = kjóla keyrir ’driver of ships’. 936 Hofuðlausn 1 ‘Ocean-going craft’ = eik. -”-”‘Head’ = *munkno˛ rr ’mind ship’. 960 Sonatorrek 3 Nokkvi? 962 Arinbjarnarkviða 21 ‘House’ = legvers kno˛ rr ’bed ship’. c. 907 Lausavísa 1 ’Warship’ = fley and kno˛ rr. 934 Lausavísa 22 ‘Craft’ = karfi. 10 c. Hákonardrápa 2 ’Warships’ = skeiðar (pl.) -”Hákonardrápa 7 ’Warships’ = snekkjur / skeiðar (pl.). Before ‘Kvad om Erik 950. blodøkse’ Comment ‘Sea’ = ferju bakki ‘hill of the ferry’. -“- Manuscripts míns knorr. Corrupt stanza. Variants in the manuscripts. Dated to before 960 by Jesch (2001a: 126, footnote 18). T  V   64 71 83 99 115 Eyvindr skáldaspillir Kormákr O˛ gmundarson -”- c. 965 Lausavísa 9 955– 970 955– 970 Gísli Súrsson 970s Lausavísa 4 Lausavísa 57 Lausavísa 15 -”- Vígfúss Víga- c. 986 Glúmsson Brúsi 2ⁿ Hallason half of 10 c. Einarr skála- c. 986 glamm -”c. 986 -”- -”- c. 986 -”- 6 124 -”- c. 986 126 Bjorn Breiðvíkingakappi Steinunn Dalksdóttir Eysteinn Valdason Þorleifr jarlsskáld -”- c. Lausavísa 5 1000 c. 999 Lausavísa 2 116 117 128 131 134 -”137 137 138 138 142 147 148 151 ‘Digt om Hakon jarl (?)’ Lausavísa Vellekla 2 -”- 3 c. ‘Et digt om Tor’ 2 1000 c. 990 Lausavise 5 -”- 139 ‘Shield’ = Ullar kjóll The god Ullr’s ship was ’Ullr’s ship’. named Skjoldr ‘shield’. ’Game board’= húnkno˛ rr ’game piece ship’. ’Some ship in the king’s fleet’ = elliði ‘Sea’ = braut fleya ‘the road of ships’. ’Warship’ = kno˛ rr. ‘Shield’ = fleygarðr ‘ship fence’. ’Shield’ = Ullar askr ’Ullr’s ship’. ’Warrior’ = orþeysir flausta ‘driver of ships’. ‘Poem’ = fley berg-Saxa ‘ship of the dwarfs’. See comment to 64. Two dwarfs saved themselves from a skerry by giving away the (mead of) poetry; it thus functioned as a ship. A warship referred to as a borðróinn barði ‘a ship rowed on both sides’ ’Atlantic-going craft’ = flaust. ’Atlantic-going craft’ = kno˛ rr. ’Two-man rowing boat’ = flaust. ’dinghy’ = bátr. -”- ’Atlantic-going cargo ship’ = kno˛ rr x 2. Tindr c. 987 ‘Drape om Hakon ’Warships’ = skeiðar Hallkellsson jarl’ 4 (pl.). -”-”-”- 5 ’Warships’ = skeiðar -”-”-”- 9 ’Warships’ = skeiðar -”-”-”- 10 ’Warships’ = skeiðr Eilífr c. Þórsdrápa 14 ’Breast’ = hlátr-elliði Goðrúnarson 1000 ’laughter ship’. c. 990 Hákonardrápa 1 ’Shield’ = Ullar askr Hallfreðr vandræðaskáld ’Ullr’s ship’. -”996 Óláfsdrápa 1 ’Warships’ = herskip. -”1001 Óláfsdrápa, erfidrápa ’Warships’ = skeiðar 6 FJ understands Elliði as a name. See comment to 64. 140 E H 153 -”- -”- -”- 14 158 -”- 996 Lausavísa 4 158 -”- c. 1000 Lausavísa 5 162 -”- Lausavísa 24 162 -”- c. 1000 -”- 163 -”- -”- Lausavísa 26 167 169 Anonymous Anonymous 10 c. Oddmjór 996 ‘Om Stefnirs skib’ 169 172 Anonymous, c. 999 Lausavísa ‘nokkvamaðr’ Anonymous 10 c. 173 Anonymous 10 c. 174 Anonymous 10 c. 176 Anonymous c. 980 Lausavísa -”- ’The warship Ormrinn -”langi ’the long serpent’’ = et langa lung ’the long ship’. ’Atlantic-going craft’ = kno˛ rr. ‘Poem’ = nokkvi Austra See comment to 117–3. burar ’ship of the dwarf ’. ‘A splendid, beautiful (war?)ship’ = fley. ‘A splendid, beautiful (war?)ship’ = skeið. ‘Atlantic-going ship’ = kno˛ rr ‘Warship’ = skeið. ’Atlantic-going ship’ = kno˛ rr. ‘Single-man rowing boat’ = nokkvi. ‘Craft in general’ = bátar (pl.). ‘Poem(s)’ = skip dverga See comment to 117–3. ‘ship of dwarfes’. ‘Sea’ = flausta ferill ‘the road of ships’. ‘Warships’ = skeiðar. . Terms for ship types in runic inscriptions before  AD Selected from Jesch a (:  ff.). Unspecified terms like skip are not mentioned. Location of inscription / name Tryggevælde, Zealand, Denmark Code Term / text Dating DK Sj 82 raknhiltr … sati stain þansi auk c. 900 karþi hauk þansi … auk skaiþ þaisi ‘Ragnhildr …placed this stone and made this mound … and this skeið (stone ship)…’ Fresta church, U 258 ....on trabu nurminr o kniri 980 Uppland, asbiarnaR ‘...The Norwegians –1015 Sweden killed him on Ásbjorn’s kno˛ rr’ Comment Refers to a stone ship setting (now lost) of which the rune stone was probably a part. Nielsen 2006: 301 Samnordisk runtextdatabas, http://www.nordiska. uu.se/forskn/samnord.htm, 12.05.2011 / Gräslund 2006: 126. T  V   141 . Terms for ship types in Eddic poetry The list has been compiled by looking up all the ship type terms mentioned by Falk () in Finnur Jónsson’s Lexicon poeticum (–), checking the attestations in Bugge’s edition () of the Eddic poems and searching for the terms in a digital text file of the Eddic poems from http://heimskringla.no [address checked  November ].¹⁵² Voluspá 51 Hávamál 74 Hávamál 82 Hávamál 154 Grímnismál 43–44 Hárbarðsljóð 7 -“- 39 -“- 53 Rígsþula 48 Hymiskviða 19 -“- 33 Volundarkviða 33 Helgakviða Hjorvarðssonar 12, 18, 19, 23, 28 Helgakviða Hundingsbana I, 23 -“-, 24 -“-, 30 -“- 31 -“-, 49 Helgakviða Hundingsbana II, 5, 6 Helgakviða Hundingsbana II, 13 Helgakviða Hundingsbana II, 19 Helgakviða Hundingsbana II, 32 Helgakviða Hjorvarðssonar 14 Sigurðarkviða in skamma 53 Guðrúnarkviða in forna 16 Atlamál 98 Atlamál 103 ¹⁵² Falk (:  ff. A kjóll carries the giants to Ragnarok. Mention of skips ráar ‘ships’ sail yards’. Mention of skip in connection with speed. ‘Some craft’ = far. ‘Freyr’s Skíðblaðnir’ = skip. ‘Rowing boat’ = eikja. ‘Ocean-going craft’ = skip. ‘Rowing boat’ = bátr. ‘A kings’ ship’ = kjóll. ‘A warship’ = kjóll. ‘Brewing container’ = olkjóll ‘ale ship’. Mention of skips planking. ‘Warships’ = skip (pl.) ‘Warships’ = beit (pl.) and skip (pl.). ‘Warships’ = skip (pl.). ‘Warships’ = far. ‘Fleet of warships’ = flaust (pl.). ‘Warships’ = kjólar (pl.). ‘Warships’ = fley (pl.). ‘Warships’ = langskip (pl.). ‘Warships’ = skip (pl.) ‘Unspecified ship’ = skip (pl.). ‘Warship’ = beit. ‘Some craft’ = far. ‘Splendid warship’ = skip. ‘Long-distance ship’ = skip. ‘Burial ship’ = kno˛ rr. 142 E H . Terms for ship types only used in poetry Beit n., eik f., flaust n., lung n., nór m., regg n. In addition, askr, elliði and kjóll have very limited use in prose; see the discussions above. The list is made by comparing Falk’s list of Old Norse terms for ship types () with Finnur Jónsson’s Lexicon Poeticum, Fritzner –, and A Dictionary of Old Norse Prose, http://dataonp. hum.ku.dk/, whose database has been very useful in other parts of the study, too.¹⁵³ Terms only occurring in Þulur (footnote ) are not included. . References: Aasen, Ivar, 1873: Norsk Ordbog. Anden forøgede Udgave af Ordbog over det norske Folkesprog. Christiania: Malling. Ahlbäck, Olav and Bertil Bonns, 1986: “Båtar och båttyper i Kvarkens båtmuseum”. Västerbotten. Västerbottens Läns Hembygdsförenings Årsbok 2–86. 98–107. Alexander Jóhannesson, 1923: Íslenzk tunga í fornöld. Reykjavík: Bókaverzlun Ársæls Árnasonar. Ásgeir Blöndal Magnússon, 1989: Íslensk orðsifjabók. Reykjavík: Orðabók Háskólans. Bergsveinn Birgisson, 2008: Inn i skaldens sinn. Kognitive, estetiske og historiske skatter i den norrøne skaldediktingen. Dr. art. thesis, University of Bergen. Bill, Jan et. al., 1998: “Nordeuropas største skibsfund. Ni vrag fra vikingetid og middelalder under museumsøen i Roskilde”. Nationalmuseets arbejdsmark. 136–66. ———, 2000: “Roskildeskibene.” In Tom Christensen and Michael Andersen (ed.): Civitas Roscald. Fra byens begyndelse. Roskilde: Roskilde Museums Forlag. 211–67. Bitterli, Dieter, 2009: Say What I Am Called: The Old English Riddles of the Exeter Book and the Anglo-Latin Riddle Tradition. Toronto: University of Toronto Press. Bjørgo, Narve, 1965: “Skipstypar i norrøne samtidssoger”. Sjøfartshistorisk årbok. 7–20. Blindheim, Martin, 1985: Graffiti in Norwegian stave churches c. 1150-c. 1350. Medieval art in Norway. Doktoravhandling, Universitetet i Oslo, 1986. Oslo: Universitetsforlaget. ¹⁵³ Falk :  ff. T  V   143 Bosworth, Joseph and T. Northcote Toller, 1898: An Anglo-Saxon dictionary. Based on the manuscript collections of the late Joseph Bosworth. Edited and enlarged by T. Northcote Toller. Oxford: Oxford University Press. Brennu-Njáls saga, 1954. Íslenzk fornrit 12. Ed. Einar Ól. Sveinsson. Reykjavík: Hið íslenzka Fornritafélag. Bugge, Sophus (ed.), 1867: Sæmundar Edda hins fróða. (Norrœn Fornkvæði). Christiania: Malling. Christensen, Arne Emil, 1982: “Viking Age Ships and Shipbuilding in Norway”. Norwegian Archaeological Review 15. 19–28. ———, 1992: “Skipet.” In Arne Emil Christensen et. al. (ed.): Osebergdronningens grav. Vår arkeologiske nasjonalskatt i nytt lys. Oslo: Schibsted. 138–53. Cordes, Gerhard et. al., 1928-: Mittelniederdeutsches Handwörterbuch. Agathe Lasch und Conrad Borchling, fortgeführt von Gerhard Cordes. Neumünster: Wachholtz. Crumlin-Pedersen, Ole, 1970: “Skibstyper. Arkæologiske vidnesbyrd.” Kulturhistorisk leksikon for nordisk middelalder 15. Oslo: Gyldendal. 482–91. ———, 1991: “Ship Types and Sizes AD 800–1400.” In Ole Crumlin-Pedersen (ed.): Aspects of maritime Scandinavia AD 200–1200. Proceedings of the Nordic Seminar on Maritime Aspects of Archaeology, Roskilde, 13th-15th March, 1989. Roskilde: The Viking Ship Museum. 69–82. ———, 1997a: “Large and small warships of the North.” In Anne Nørgård Jørgensen and Birthe L. Clausen (ed.): Military aspects of Scandinavian society in a European perspective, AD 1–1300. Papers from an international research seminar at the Danish National Museum, Copenhagen, 2–4 May 1996. Publications from the National Museum. Studies in archaeology & history, 2. Copenhagen: The National Museum. 184–94. ———, 1997b: Viking-Age ships and shipbuilding in Hedeby/ Haithabu and Schleswig. Ships and boats of the North 2. With contributions by Christian Hirte, Kenn Jensen and Susan Möller-Wiering. Schleswig: Archäologisches Landesmuseum der Christian-Albrechts-Universität. Crumlin-Pedersen, Ole and Olaf Olsen (ed.), 2002: The Skuldelev ships I. Ships and boats of the North 4. Roskilde: The Viking Ship Museum / National Museum of Denmark / Centre for Maritime Archaeology. Diplomatarium Norvegicum, 1847–1995. 22 volumes. Christiania / Oslo: Malling / Norsk historisk Kjeldeskrift-Institutt. 144 E H Edda Snorra Sturlusonar, 1931. Udgivet efter håndskrifterne af Kommissionen for Det Arnamagnæanske Legat ved Finnur Jónsson. København: Gyldendal. Egils saga Skalla-Grímssonar, 1933. Íslenzk fornrit 2. Ed. Sigurður Nordal. Reykjavík: Hið íslenzka fornritafélag. Eldjarn, Gunnar, 1990: Nordlandsbåten. Nordlandsbåten og Åfjordsbåten 2. Rissa: Båtstikka D/A. Ellmers, Detlev, 1972: Frühmittelalterliche Handelsschiffahrt in Mittel- und Nordeuropa. Offa-Bücher 28. Neumünster: Wachholtz. ———, 2004: “Schiff und Schiffsarten.” In H. Arbman et. al. (ed.): Reallexikon der Germanischen Altertumskunde 27. Begründet von Johannes Hoops. Herausgegeben von Heinrich Beck [et al.]. Berlin: de Gruyter. 13–20. Engelhardt, Conrad, 1865: Nydam mosefund 1859–1863. Udgivet med understøttelse af det kongellige ministerium for hertugdømmet Slesvig. Kjøbenhavn. Evans, A. C. and R. Bruce-Mitford, 1975: “The Ship.” In Rupert Bruce-Mitford (ed.): The Sutton Hoo ship-burial 1. Excavations, background, the ship, dating and inventory. London: British Museum Publications. 345–435. Falk, Hjalmar, 1912: “Altnordisches Seewesen”. Wörter und Sachen 4. 1–121. Falk, Hjalmar and Alf Torp, 1903–06: Etymologisk ordbog over det norske og det danske sprog. Kristiania: Aschehoug. Felbo, Mette, 1999: Skibsbilleder i Skandinavien 800–1400. Kontekst og funktion, indhold og kildeverdi. 3 volumes. Cand. phil.-speciale, Institut for Arkæologi og Etnologi, Københavns Universitet. Fidjestøl, Bjarne and Odd Einar Haugen, 1999: The dating of Eddic poetry. A historical survey and methodological investigation. Bibliotheca Arnamagnæana 41. Edited by Odd Einar Haugen. Copenhagen: Reitzel. Finnbogi Guðmundsson, 1965: Orkneyinga saga. Íslenzk fornrit 34. Edition with comments. Reykjavík: Hið Íslenska fornritafélag. Finnur Jónsson, 1912–15: Den Norsk-islandske skjaldedigtning. 4 volumes. Band A 1 og A 2 tekst etter handskriftene, B 1 og B 2 restituert tekst. København: Gyldendal. ———, 1913–16: Lexicon poeticum. Forfattet af Sveinbjörn Egilsson. Forøget og påny udgivet for Det kongelige Nordiske Oldskriftselskab ved Finnur Jónsson. København: S. L. Møllers Bogtrykkeri. T  V   145 Flateyjarbok, 1860–68. 3 volumes. Udgiven efter offentlig Foranstaltning [af Gudbrandr Vigfusson og C.R. Unger]. Christiania: Malling. Foote, Peter, 1978: “Wrecks and Rhymes.” In Thorsten Andersson and Karl Inge Sandred (ed.): The Vikings. Proceedings of the Symposium of the Faculty of Arts of Uppsala University, June 6–9, 1977. Symposia Universitatis Upsaliensis annum quingentesimum celebrantis 8. Stockholm: Almqvist & Wiksell. 222–35. Friðþjófs saga ins frækna, 1954. In Fornaldar sögur Norðurlanda III. Ed. Guðni Jónsson. [Reykjavík?]: Íslendingasagnaútgáfan. 75–104. Fritzner, Johan, 1883–96: Ordbog over Det gamle norske Sprog 1–3. 2ⁿ ed. 1 ed. Christiania 1867. Kristiania: Den norske Forlagsforening. Fulk, R.D et. al. (ed.), 2008: Klaeber’s Beowulf and the fight at Finnsburg. Based on the third edition with first and second supplements of Bowulf and the Fight at Finnsburg, edited by Fr. Klaeber, Lexington, Massachusets, 1950. Toronto: University of Toronto Press. Færøyvik, Øystein, 1987: Vestlandsbåtar. Frå oselvar til sunnmørsåttring. Norske båtar 5. Oslo: Grøndahl & søn. Gallé, Johan Hendrik, 1903: Vorstudien zu einem altniederdeutschen Wörterbuch. Leiden: Brill. Gjessing, Helge, 1925: “Vest-Agder i forhistorisk tid.” In G. Midttun (ed.): Vest-Agder. Norske bygder II. Bergen: John Grieg. 33–75. Godal, Jon and Gunnar Eldjarn, 1990: System og oversyn. Nordlandsbåten og Åfjordsbåten IV. Rissa: Båtstikka. Greenhill, Basil, 2000: “The Mysterious Hulc”. The Mariner’s Mirror 86. 3–18. Grimm, Jacob and Wilhelm Grimm, 1854–1961: Deutsches Wörterbuch. 32 volumes. Leipzig: Hirzel. Grimm, Oliver, 2001: “Norwegian boathouses from the late Roman and Migration Periods.” In Birger Storgaard (ed.): Military aspects of the aristocracy in Barbaricum in the Roman and Early Migration Period. Papers from an International Research Seminar at the Danish National Museum, Copenhagen, 10–11 December 1999. Publications from the National Museum. Studies in archaeology and history 5. København: National Museum. 55–66. Gräslund, Anne-Sofie, 2006: “Dating the Swedish Viking-Age rune stones on stylistic grounds.” In Marie Stoklund et. al. 146 E H (ed.): Runes and their secrets. Studies in runology. Copenhagen: Museum Tusculanum. 117–39. Gøthche, Morten, 2000: Rekonstruktion af Nydambåden. Nationalmuseets Marinarkæologiske Undersøgelser. Rapport. ———, 2006: “The Roskilde Ships.” In Lucy Blue et. al. (ed.): Connected by the sea. Proceedings of the Tenth International Symposium on Boat and Ship Archaeology, Roskilde 2003. International Symposium on Boat and Ship Archaeology (ISBSA) 10, Roskilde, 2003. Oxford: Oxbow Books. 252–58. Hákonar saga Ívarssonar, 1952. Samfund til Udgivelse af Gammel Nordisk Litteratur, Skrifter 62. Udgivet for Samfund til Udgivelse af Gammel Nordisk Litteratur ved Jón Helgason og Jakob Benediktsson. København: Samfundet. Hammershaimb, V. U., 1891: Færøsk anthologi. 2 volumes. Samfund til udgivelse af Gammel nordisk litteratur, Skrifter 15. København: Møller. Hausen, Jürgen, 2000: “Shipping between Danube and Rhine from Medieval to modern times (canal of Charlemagne, canal of king Louis of Bavaria, Rhine-Danube-canal).” In Jerzy Litwin (ed.): Down the river to the sea. Proceedings of the Eight International Symposium on Boat and Ship Archaeology Gdansk 1997. Gdansk: Polish Maritime Museum. 21–26. Heide, Eldar, 2003: “Stamnkarakteriserande skips- og båtnemningar. Knorr og geitbåt”. Norsk sjøfartsmuseum. Årbok 2002. 57–72. ———, 2009: “More inroads to pre-Christian notions, after all? The potential of late evidence.” Á austrvega. Saga and East Scandinavia. Preprint Papers of The 14 International Saga Conference. Uppsala, 9–15 August 2009. Online at http:// www.saga.nordiska.uu.se/preprint/. 361–68. ———, 2012: “Loki, the Vätte and the Ash Lad. A study combining Old Scandinavian and late material”. Viking and Medieval Scandinavia 2011. 63–106. ———, manuskript 2012: “The etymology of cēol / kiol / kjóll ‘an Early Germanic ship type’”. Manuscript, 8 pages. Heide, Eldar and Terje Planke, manuscript 2012: “Viking ships with angular stems – Old Norse beit?”. Manuscript, 19 pages. Heimskringla, 1941. 3 volumes. Íslenzk fornrit 26–28. By Snorri Sturluson. Ed. Bjarni Aðalbjarnarson. Reykjavík: Hið íslenzka fornritafélag. Hervarar saga ok Heiðreks, 1954. In Fornaldar sögur Norðurlanda II. Ed. Guðni Jónsson. [Reykjavík?]: Íslendingasagnaútgáfan. 1–71. T  V   147 Hjardar, Kim and Vegard Vike, 2011: Vikinger i krig. Oslo: Spartacus. Holthausen, F., 1954: Altsächsisches Wörterbuch. Niederdeutsche Studien. Münster: Böhlau. Hustad, Tom and Asbjørn Klepp, 1982: “”Fembøring” = “rask (liten) føringsbåt”?”. Håløygminne. 225–43. Hversu Noregr byggðist, 1954. In Fornaldar sögur Norðurlanda II. Ed. Guðni Jónsson. [Reykjavík?]: Íslendingasagnaútgáfan. 75–87. Höckmann, Olaf, 1993: “Late Roman Rhine vessels from Mainz, Germany”. The International Journal of Nautical Archaeology 22.2. 125–35. Hødnebø, Finn, 1972: Ordbog over Det gamle norske Sprog IV. Rettelser og tillegg til J. Fritzner 1883–96: Ordbog over det gamle norske Sprog I-III. Oslo: Universitetsforlaget. Jesch, Judith, 2001a: Ships and men in the late Viking Age. The vocabulary of runic inscriptions and skaldic verse. Woodbridge: Boydell Press. ———, 2001b: “Women and ships in the Viking world”. Northern Studies 36. 49–68. Karg-Gasterstädt, Elisabeth and Theodor Frings, 1952-: Althochdeutsches Wörterbuch. Auf Grund der von Elias von Steinmeyer hinterlassenen Sammlungen im Auftrag der Sächsischen Akademie der Wissenschaften zu Leipzig, begründet von Elisabeth Karg-Gasterstädt und Theodor Frings; herausgegeben von Rudolf Grosse. 5 volumes so far. Berlin: Akademie-Verlag. Klepp, Asbjørn, 1983: Nordlandsbåter og båter fra Trøndelag. Norske båter 4. [Oslo?]: Grøndahl & søn. Kluge, Friedrich and Elmar Seebold, 2002: Etymologisches Wörterbuch der deutschen Sprache. 24., durchgesehene und erweiterte Auflage. Bearbeitet von Elmar Seebold. Berlin / New York: De Gruyter. Korhonen, Olavi, 1982: Samisk-finska båttermer och ortnamnselement och deras slaviska bakgrund en studie i mellanspråklig ordgeografi och mellanfolklig kulturhistoria. Skrifter utgivna av Dialekt-, ortnamns- och folkminnesarkivet i Umeå. Serie A, Dialekter 3. Doktoravhandling, Universitetet i Uppsala, 1982. Umeå. ———, 1987: “Öst-västliga kulturkontakter i belysning av en båtterm (karv / karfi/ karbas etc.).” In Guy Björklund (ed.): Bottnisk kontakt III. Föredrag hållna vid maritimhistoriskt seminarium i Jakobstad 7.-9.2.1986. [Jakobstad]. 11–21. Krag, Claus, 1995: Vikingtid og rikssamling. 800–1130. Asche- 148 E H hougs Norgeshistorie 2. Edited by Knut Helle. Oslo: Aschehoug. ———, 2000: Norges historie fram til 1319. Oslo: Universitetsforlaget. Krapp, George Philip, 1932: The Vercelli book. The Anglo-Saxon poetic records: A collective edition, 2. New York: Columbia University Press. Kruyskamp, C. , 1961: Groot woordenboek der Nederlandse taal. Achtste, geheel opnieuw bewerkte en zeer vermeerderde druk. ‘s-Gravenhage: Martinus Nijhoff. Kuhn, Hans, 1973: “Asch.” In H. Arbman et. al. (ed.): Reallexikon der germanischen Altertumskunde I. Berlin: de Gruyter. 449–50. ———, 1983: Das Dróttkvætt. Heidelberg: Carl Winter. Landnámabók, 1968. In Íslendingabók. Landnámabók. Íslenzk fornrit 1. Ed. Jakob Benediktsson. Reykjavík: Hið íslenzka fornritafélag. 29–397. Lassen, Annette, 2011: Odin på kristent pergament. En teksthistorisk studie. Annette Lasse n. København: Museuem Tusculanum. Lindqvist, Sune, 1941–42: Gotlands Bildsteine. 2 volumes. Kungl. Vitterhets historie och antikvitets akademien. Archäologische Monographien. Gesammelt und untersucht von Gabriel Gustafson und Fredrik Nordin mit Zeichnungen von Olof Sörling. Photographien von Harald Faith-Ell. Nach erneuter Durchsicht und Ergänzung des Materials herausgegeben von Sune Lundqvist. Stockholm: Wahlström & Widstrand. Livingston, Michael, 2011: The Battle of Brunanburh: A Casebook. Exeter Medieval Texts and Studies. Exeter: University of Exeter Press. Low, George, 1879: A tour through the islands of Orkney and Schetland. Containing hints relative to their ancient modern and natural history collected in 1774. Kirkwall: William Peace & Son. Malmros, Rikke, 1985: “Leding og skjaldekvad. Det elvte århundredes nordiske krigsflåder, deres teknologi og organisation og deres placering i samfundet, belyst gennem den samtidige fyrstedigtning”. Aarbøger for nordisk Oldkyndighed og Historie. S. 89–139. ———, 2010: Vikingernes syn på militær og samfund. Belyst gennem skjaldenes fyrstedigtning. Århus: Aarhus Universitetsforlag. McCusker, John J., 1966: “The Wine Prise and Medieval Mercantile Shipping”. Speculum 41–2. 279–96. Meulengracht Sørensen, Preben, 1991: “Om eddadigtenes alder.” In Gro Steinsland et. al. (ed.): Nordisk hedendom. Et sympo- T  V   149 sium. Odense: Odense Universitetsforlag. 217–28. Moltke, Erik, 1985: Runes and their origin. Denmark and elsewhere. Updated and rev. [ed.]. Translated by Peter G. Foote. Original title: Runerne i Danmark og deres oprindelse. Copenhagen: Nationalmuseets Forlag. Myhre, Bjørn, 1976: “Nausttuft fra eldre jernalder på Stend i Fana. Nausttuftene som kilde til kunnskap om bygningskonstruksjon, bosetning og samfunnsforhold.”. Viking. 29–79. Myrvoll, Klaus Johan, 2009: Samstöfur seinar eda skjótar. Ein etterrøknad av trykk- og kvantitetstilhøve i skaldeversemålet dróttkvætt. Masteroppgave i norrøn filologi, Universitetet i Oslo, 2009. Nicolaysen, N., 1882: Langskibet fra Gokstad ved Sandefjord. The Viking-Ship from Gokstad. Kristiania: Cammermeyer. Nielsen, M. L., 2006: “Tryggevælde.” In Rosemarie Müller et. al. (ed.): Reallexikon der Germanischen Altertumskunde 31. Begründet von Johannes Hoops. Herausgegeben von Heinrich Beck [et al.]. Berlin: de Gruyter. 300–03. Nordeide, Sæbjørg Walaker, 1989: “...De beste bønder i kiøbstæden...”. En funksjons- og aktivitetsanalyse basert på gjenstandsmaterialet. Meddelelser nr. 20 fra prosjektet Fortiden i Trondheim bygrunn: Folkebibliotekstomten. Trondheim: Riksantikvaren, Utgravningskontoret for Trondheim. Norges gamle Love indtil 1387, 1846–95. 5 volumes. Ed. R. Keyser og P. A. Munch. Christiania: Chr. Grøndahl. Norsk Ordbok, 1966-. Edited by Alf Hellevik and Lars S. Vikør. Ordbok over det norske folkemålet og det nynorske skriftmålet. Oslo: Det Norske Samlaget. Nylén, Erik and Jan Peder Lamm, 1987: Bildstenar. 2. omarbetade och kompletterade upplaga. Stockholm: Gidlund. Ólafur Halldórsson, 1978: Grænland i miðaldaritum. Doctoral thesis, University of Iceland: Sögufélag. Olsen, Magnus, 1931: “Gammelnorske skibsnavn”. Maal og Minne. 33–43. Ordbog over det danske sprog, 1918–56. 28 volumes. Udgivet af det Danske Sprog- og Litteraturselskab. København: Gyldendal. Ordbok över svenska språket, 1898 -. Utgiven av Svenska akademien. [Ordbok öfver svenska språket / Svenska akademiens ordbok.]. Lund: Gleerup. Orkneyinga saga, 1965. Íslenzk fornrit 34. Ed. Finnbogi Guðmundsson. Reykjavík: Hið Íslenska fornritafélag. Osborn, Marijane, 2003: “Norse Ships at Maldon: The Cultural 150 E H Cotext of Æschere in the Old English Poem ‘The Battle of Maldon’”. Neuphilologische Mitteilungen 104. 261–80. Past, E., 1946: “Lodjan – en urgammal segelfartygstyp i Estland”. Unda Maris 1945. 89–95. Pedersen, Walter Meyer, 2002: Bårsetbåten. En revurdering av rekonstruksjonen fra 1937. Hovedfagsavhandling i arkeologi, Universitetet i Tromsø, 2002. Tromsø. Refsum, Helge, 1933: “”Nu” og andre glømte farkoster på Romerike”. Maal og Minne. 74–76. Rendboe, Laurits, 1987: Det gamle shetlandske sprog. George Low’s ordliste fra 1774. North-Western European Language Evolution, Supplement 3. [Odense]: Odense Universitetsforlag. Rieck, Flemming, 1994: Jernalderkrigernes skibe. Nye og gamle udgravninger i Nydam Mose. Roskilde: Vikingeskibshallen. Rygh, O. and C. F. Lindberg, 1885: Norske Oldsager. Ordnede og forklarede af O. Rygh, tegnede paa Træ af C.F. Lindberg. Christiania: Cammermeyer. Sandnes, Jørn and Kaisa Rautio Helander, 1997: “Namsen.” In Jørn Sandnes and Ola Stemshaug (ed.): Norsk stadnamnleksikon. Oslo: Det Norske Samlaget. 324. Sayers, William, 1996: “The Etymology and Semantics of Old Norse knörr cargo ship. The Irish and English Evidence”. Scandinavian Studies 68. 279–90. ———, 2006: “Æschere in The Battle of Maldon: Fleet, Warships’ Crews, Spearmen, or Oarsmen?”. Neuphilologische Mitteilungen 107. 199–205. Schiller, Karl and August Lübben, 1875: Mittelniederdeutsches Wörterbuch. Bremen. Schnall, Jens Eike, 2004: “Homlubarði – ein ruderloses Schiff, auf Grund gesetzt. Zu Konungs skuggsjá, Barlaams saga ok Josaphats und skandinavischen Ortsnamen.” In Astrid van Nahl et. al. (ed.): Namenwelten. Orts- und Personennamen in historischer Sicht. 277–91. Shetelig, Haakon and Fr. Johannessen, 1929: Kvalsundfundet og andre norske myrfund av fartøier. Bergens museums skrifter. Ny række II:2. Med botaniske bidrag av Jens Holmboe og Knud Jessen. Bergen: Bergens museum. Shetelig, Haakon and Hjalmar Falk, 1937: Scandinavian archaeology. Translated by E.V. Gordon. New York: Hacker Art Books. Sigfús Blöndal, 1920: Islandsk-dansk Ordbog. Íslensk-dönsk orðabók. Reykjavík: I kommission hos Verslun Þórarins B. Þorlákssonar. Simek, Rudolf, 2005: “Skeið.” In Rosemarie Müller et. al. (ed.): T  V   151 Reallexikon der Germanischen Altertumskunde 29. Begründet von Johannes Hoops. Herausgegeben von Heinrich Beck [et al.]. Berlin: de Gruyter. 635. ———, 2006: Lexikon der germanischen Mythologie. Kröners Taschenausgabe Band 368. Dritte, völlig überarbeitete Auflage. Stuttgart: Kröner. Simek, Rudolf and Hermann Pálsson, 1987: Lexikon der altnordischen Literatur. Kröners Taschenausgabe 490. Stuttgart: Kröner. Sorokin, Petr, 2000: “The Medieval Boatbuilding Tradition of Russia.” In Jerzy Litwin (ed.): Down the river to the sea. Proceedings of the Eight International Symposium on Boat and Ship Archaeology Gdansk 1997. Gdansk: Polish Maritime Museum. 37–44. Sorokin, Petr E. , 1997: Wodnie puti i sudostroenie na severo-zapade Rusi w srednevekovie. St. Petersburg. Steen, Sverre, 1934: “Fartøier i Norden i middelalderen.” In Adolf Schück and Johannes Brøndsted (ed.): Handel og samfærdsel i oldtiden, A. Handel och samfärdsel under medeltiden, B. Nordisk kultur 16. Stockholm: Bonnier. 282–300. Stylegar, Frans-Arne and Oliver Grimm, 2005: “Boathouses in Northern Europe and the North Atlantic”. The International Journal of Nautical Archaeology 34.2. 253–68. Summarium Heinrici. Band 1. Textkritische Ausgabe der ersten Fassung. Buch I-X, 1974. Quellen und Forschungen zur Sprachund Kulturgeschichte der germanischen Völker Neue Folge. Herausgegeben von Reiner Hildebrandt. Berlin: de Gruyter. Svavar Sigmundsson, 2008: “Elliðaár”. Published online by The Árni Magnússon Institute for Icelandic Studies in the series Örnefni mánaðarins, at http://www3.hi.is/Apps/WebObjects/ HI.woa/1/wa/dp?id=1023962&wosid=KuMPbMqLZ2t9zP W8LSNwc0. Sverris saga, 2007. Íslenzk fornrit 30. Ed. Þorleifur Hauksson. Reykjavík: Hið íslenska fornritafélag. Sävborg, Daniel, 2004: “Om eddadikternas ursprung och ålder. Gamla och nya tankar”. Arkiv för nordisk filologi 129. 55–104. Söderwall, K. F., 1884–1918: Ordbok öfver svenska medeltidsspråket. 3 volumes. Lund: Berlingska boktryckeri. Sørensen, Anne C., 2001: Ladby. A Danish ship-grave from the Viking Age. Ships and boats of the North 3. Roskilde: The Viking Ship Museum. Sörla þáttr, 1860. In Flateyjarbók I. Udgiven efter offentlig Foranstaltning [af Gudbrandr Vigusson og C. R. Unger]. Christiania: Malling. 275–83. 152 E H Tanaka, Sofia, 2010: “Unikt vrakfynd i Stockholm”. Dagens nyheter. Thier, Katrin, 2002: Altenglische Terminologie für Schiffe und Schiffsteile. Archäologie und Sprachgeschichte 500–1100. BAR international series 1036. Oxford: Archaeopress. Thowsen, Atle, 1970: “The Norwegian Export of Boats to Shetland, and its Influence upon Shetland Boatbuilding and Usage”. Sjøfartshistorisk årbok 1969. 145–208. Two of the Saxon chronicles parallel. With supplementary extracts from the others. A. revised text, 1892. Edited with introduction, notes, appendices, and glossary by Charles Plummer on the basis of an edition by John Earle. Oxford: Clarendon Press. Varenius, Björn, 1992: Det nordiska skeppet. Teknologi och samhällsstrategi i vikingatid och medeltid. Stockholm studies in archaeology 10. Doctoral thesis, University of Stockholm, 1992. Vries, Jan de, 1962: Altnordisches etymologisches Wörterbuch. Zweite verbesserte Auflage. 1 ed. 1961. Leiden: E.J. Brill. Westerdahl, Christer, 1985a: “Sewn boats of the North. A preliminary catalogue with introductory comments. Part 1”. The International Journal of Nautical Archaeology and Underwater Exploration 14.1. 33–62. ———, 1985b: “Sewn boats of the North. A preliminary catalogue with introductory comments. Part 2”. The International Journal of Nautical Archaeology and Underwater Exploration 14.2. 119–42. Whaley, Diana (ed.), 2012: Poetry from the Kings’ Sagas 1: From Mythical Times to c. 1035. Skaldic poetry of the Scandinavian Middle Ages 1. Turnhout: Brepols. Whitelock, Dorothy et. al., 1961: The Anglo-Saxon Chronicle. A revised translation. London: Eyre and Spottiswoode. Williams, Hugh (ed.), 1899: Gildae De excidio Britanniae, Fragmenta, Liber de paenitentia, accedit et Lorica Gildae = Guildas: The ruin of Britain, Fragments from lost letters, The penitential, together with the Lorica of Gildas London: The Honourable Society of Cymmrodorion. Äldre Västgötalagen. I diplomatariskt avtryck och normaliserad text jämte inledning och kommentar, 1919. Skrifter utgivna av Svenska litteratursällskapet i Finland 144. Ed. Bruno Sjöros. Helsingfors: Svenska litteratursällskapet i Finland. Örvar-Odds saga, 1954. In Fornaldar sögur Norðurlanda II. Ed. Guðni Jónsson. Reykjavík: Íslendingasagnaútgáfan. 199–363. Åkerlund, Harald, 1963: Nydamskeppen. En studie i tidig skandinavisk skeppsbygnadskonst. Göteborg: Sjöfartsmuseet. T  V   153 Samandrag Dei skipstypane vi kan lesa om i Heimskringla og resten av den norrøne litteraturen – korleis såg dei ut? Er det mogeleg å kople dei til skipsfunna og skipsbileta som vi kjenner frå vikingtida og mellomalderen? Denne artikkelen er eit forsøk på å gjera eit slik kopling når det gjeld tidleg vikingtid. Eit grunnleggjande kjeldeproblem er at den norrøne prosaen, som er frå –talet, skildrar den skipsteknologien forfattarane kjende frå si samtid, også når dei fortel om hendingar mange hundre år tidlegare, då andre skipstypar og andre tekniske løysingar var i bruk. Likevel inneheld det norrøne tekstkorpuset nemningar på eldre skipstypar og sporadiske opplysningar om dei. Forfattaren prøver å sile ut det eldste laget av opplysningar, med hjelp av skaldedikt traderte i lang tid fram til dei vart nedskrivne, runeinnskrifter, utanlandske tekstar, poetiske nemningar, etymologi, jamføring med båttypar kjende frå seinare tradisjon, m.m. Han kjem fram til at knorr, beit, skeið, kjóll, askr og elliði var dei viktigaste skipstypane i Skandinavia tidleg i vikingtida, i alle fall i vest, og at nemninga knorr på den tida vart brukt om krigsskip eller kombiskip som Oseberg og fyrst seinare om frakteskip som Skuldelev . ‘Skip med attoverbøygd stamn’ ser ut til å ha vore den opphavlege tydinga til knorr. Kjólar ser ut til å ha vore dryge allroundskip som Gokstad, og beit slike skip med trekanta utfyllingar under kvar stamn som vi kjenner frå svært tidlege skipsbilete. Skeiðar var etter alt å døme lange og slanke, representerte av Ladbyskipet. Askar er ein annan svært tidleg type, og nemninga kan vise til at borda på den typen var sydde saman, som på ein sveipask. Elliðar ser ut til å ha vore ein kombinert elve- og sjøgåande type, opphavleg austeuropeisk.