The origins of runes

ROMAN CONNECTIONS

 All runic finds from the Danish bogs and graves, approximately dating from the period 160- 450, have been found in a context that clearly shows Roman connections. The bog-deposits contain Roman goods, as do the graves. Runic finds thus emerged either from a military context or a luxurious, aristocratic, context. In both cases the objects were prestige goods. The runes on the bogfinds were carved on objects that may be linked to the top of the military hierarchy (Ilkjær 1996a:70). It appears that Germanic weapons were inscribed in a similar way as Roman weapons (Rix 1992:430-432). At the time of the Marcomanni wars (161-175), contacts were established between the area of the Lower Elbe and the area of the Marcomanni. An elite from the Lower Elbe region migrated southwards and settled in the Marcomanni region (Lund Hansen 1995:390). The Danish elite from that same period must be seen in relation to Germanic vassal kings, who were, highly Romanized, living near the limes of Upper Germany/Raetia (Lund Hansen 1995:390), the region of the Marcomanni, Quadi and Iuthungi. The presence of Ringknauf swords in a warrior grave on Jutland and in deposits of the Vimose bog indicates that there were contacts with Central Europe. These second century swords are typically provincial- Roman products, and the owners, like the man from the Juttish grave of Brokær, must have taken part in the Marcomanni wars. The swords in the Vimose bog belonged to attackers from the South. The sites where these swords were found show that the route was from the Danube northwards along the Elbe (thus crossing the region of Harii and Lugii). At the same time Himlingøje (Sealand) emerged as a power-centre. Here, silver bocals with depictions of warriors holding Ringknauf swords point to the connection with the Marcomanni region (Lund Hansen 1995:386ff.). Ilkjær (1996b:457) mentions the princely grave from Gommern (Altmark, near Magdeburg, the region of the -leben placenames), which, although about a century younger, can be seen as 48 a parallel to the rich Illerup deposits. Parallels can also be detected between deposits in the Vimose and Illerup bogs concerning the collections of silver shield-buckle fragments, the pressed foil ornamentation, face-masks, weapons and military equipment. These objects mark the high military rank of the owners. Outstanding silver shield-accessories emphasize the extraordinary rank of the Germanic elite. The same custom can also be observed in lateantique Gallia, in the warrior grave of Vermand, who, by the look of his shield-accessories, was a Germanic princeps in Roman service (Ilkjær 1996b:475). Among the Illerup material of bronze and iron shield-buckles, Ilkjær notices parallels with finds from Vimose and gravegoods from Norwegian graves (Ilkjær 1996b:475). These belonged to warriors of a lower standing. An analysis of the pressed foil ornaments on the silver shields proves the close connection; the shields must have been produced in the same workshop, by Niþijo, according to Ilkjær (1996b:475). Shield-accessories like these can only be found in excessively rich graves, such as those from Gommern (Germany), Musov (Czechia), Avaldsnes (Norway) and Lilla Harg (Sweden). Therefore, the Prachtschilde from Illerup represent the very top of the elite (Ilkjær 1996b:476). He assumes this elite conducted the trade in Roman military goods (Ilkjær 1996b:477). Without these Roman goods, the extensive wars that preceded the huge offerings in the bogs, would not have been possible. The elite that organised these wars proliferated themselves by 'barbarizing’ the Roman equipment and by decorating them in a Germanic way, which was done in Germanic workshops (Ilkjær 1996b:478). Thus, although the goods make a thoroughly Roman impression, the ornamentation is indigenous, producing a splendid combination of Roman and Germanic culture. Laguþewa was one of the leading princes, according to Ilkjær (1996b:485), because of his shield with gilt-silver pressed foil and precious stones; a rich horse's garment probably belonged to him as well. Wagnijo and Niþijo were war-leaders, too, according to Ilkjær (1996:485), a statement I cannot agree with, since they were most probably weaponsmiths. The runes on several bog finds are not only found on the most precious objects, but also on humbler things such as the wooden handle for a fire-iron (Illerup V) and the comb (Vimose V).

The inscriptions on the lanceheads can directly be connected with the elite, since they controlled the production of these weapons (Ilkjær 1996b:481). From analyses of the pressed foil and pearl-wire ornamentations, it was concluded, on the basis of their highly artistically uniform nature, that there must have been extensive communication with jewellers in Central Europe. The quality of the Thorsberg finds, for instance, points to strong Roman influence. This influence is shown by the use of certain precious stones and the use of mercury (Ilkjær 1996b:481f.).

 In the meantime, in the Danish areas of eastern Sealand and Funen wealth and power accumulated and the possession of gold and silver coins increased. Roman luxury goods were imported, probably over sea via the Lower Rhine, through the Vlie along the North Sea coast, through the Limfjord and so on to the north coast of Sealand (Lund Hansen 1995:389, 408f. and the map on page 388). The commissioners who had sent for the luxury goods knew what they wanted; it was no matter of mere chance what came to the North. This also points to close contacts between the clients in the North and the elites living on the border with the Empire. During the second century, tension grew in the North Sea regions, because of pirate raids by the Chauci. One wonders how safe the route by sea-way really was, but perhaps there were treaties between the Sealand aristocrats and Chauci (and Fresones?), who controlled the North Sea coast.

 Most probably there was a relation between political events at the borders of the Roman Empire and several weapon-offerings in South Scandinavia (Ilkjær 1996b:339). The first big attack on South Scandinavia coincides with the Marcomanni wars. The offerings in the Vimose bog (Funen), of which the harja comb formed part, were contemporaneous. The attack on Funen came from the South. Further offerings in Vimose and Illerup of around 200 AD coincide with Germanic attacks on the limes. Now the attackers came from the North, from across the Kattegat. All over Scandinavia, many graves are found that contain a similar inventory of weapons. These graves are contemporaneous with the fall of the limes in the 3rd c. This was no coincidence, according to Ilkjær (1996b:339). The initial period of manufacturing weapons on a large scale was at about 200 AD, coinciding with the organisation of armies consisting of hundreds of warriors. We may suppose there existed a powerful and structural organisation at the time. The aim was not merely raiding for loot, there must have been a real struggle for power (Ilkjær 1996b:337ff.) Among the goods in the Illerup bog was an enormous amount of Roman equipment; this of course could not originate from Scandinavia. The wars, predominantly on Jutland, were fought between Scandinavians. All swords were Roman imports and may be interpreted as evidence for the existence of connections between Scandinavia and the Rhineland, according to Ilkjær (in a letter dd. 16 December 1996).

TO SUM UP

in the 2nd c., Germanic groups from the Lower Elbe region moved South, due to the Marcomanni wars in the region north of the Danube. Van Es mentions the Langobardi and the Goths who moved from regions near the Lower Elbe, the Lower Oder and Weichsel respectively (Van Es 1967:537). At the same time an attack was launched upon Denmark from southerly, continental, regions. Booty from these wars was deposited in the Vimose and Thorsberg bogs. Apparently these southerly attackers had contacts with tribes from Sealand (Lund Hansen 1995:406), which may have had something to do with a conflict between Sealand and Funen.

The alliance between Sealand and continental Germanic tribes may also explain the route of import goods: via the Rhine estuary and the North Sea, since the route over land and via the Baltic will not have been safe. In this way the route (of the propagation) of the runes can also be explored. There were contacts between the Rhine region and the North. One must assume the existence of alliances between several Scandinavian elites and continental Germanic ones, living along the Rhine- (and Danube-) limes, in the region between lower Elbe and Rhine, and south of the Baltic. The intermediaries of certain crafts and knowledge must have been individuals. Ilkjær locates Wagnijo, Niþijo's workshop and Laguþewa somewhere in the south of Norway. They belonged to a political alliance of peoples from several regions along the coast and inland valleys, according to Ilkjær (personal communication). This does not exclude the fact that they may have come from elsewhere, from the Continent. Their coming to the North may have been the result of the weapon trade between the Rhineland and Scandinavia. They belonged to the top of the military elite, as was stated by Ilkjær (see above), and it was the elite that controlled weapon import and weapon production.

A chronology of the origin of runic objects (from major find-complexes) may illustrate these contacts: The enormous weapon export to the northern barbarians may have been the result of a Roman divide-and-rule 35 policy, in order to let the Germanic tribes fight among themselves to satisfy their land-hunger. The wealth of some leaders may have been based on relations with high-placed persons in Rome. The gift-exchange system of precious objects belongs also to this atmosphere. Roman soldiers were not allowed to own their weapons - they were stateproperty. Contrary to this, Germanic mercenaries did own their weapons. Yet, very few weapons have been found in graves; apparently a weapon was a heirloom that stayed on in a family for generations. Captured weapons were dedicated to the gods and deposited in bogs. 50 1. Vimose, Funen, ca. 160 AD, from the South. 2. Thorsberg, Schleswig-Holstein ca. 200, from the South. 3. Illerup, Jutland, ca. 200-250, from the North (but made by southern weaponsmiths!)

Sealand, Jutland, Skåne, gravefinds, 200-275, luxury goods, indigenous.

 The gravecontexts, though, were Roman. The runic brooches (of nr. 4) are indigenous, so we may assume the inscriptions were made on the spot. Even here the contacts with continental Germanic tribes may also have played a role. The greater part of the names on the brooches appear to be West Germanic: hariso, lamo, alugod, maybe also widuhudaz (Makaev 1996:63). The Danish armies and the enemy from across the sea, from Sweden and Norway and from North-West Germany, fought each other with the same Roman weapons35. It is not unlikely that this was stimulated by Roman diplomacy. It is a well-known fact that the Romans donated subsidies and privileges to barbarian leaders, the foederati, to keep them in power - with the intrinsic purpose to keep them under control. In exchange for money and goods, the allied Germanic leader had to keep other barbarians away from the borders of the Empire, in order to create a bufferzone. Wars were preferably fought far away from Rome, far away from the limes and without Roman troops (Braund 1989:14-26). It appears that the knowledge of the production of strong iron weapons was not very widespread among the Germanic tribes (Much 1959:84ff.). This probably prompted the import (or the robbing) of Roman swords. Lønstrup (1988:95ff.) states that over 100 Roman swords have been found in the Illerup bog. One part carries stamps and other Roman markings, the other part has no marks, but both typologically and technologically it equals the first part; therefore these were also made in the Empire. These swords may have been bought, captured or obtained as a gift. This last possibility only applies to Germanic foederati near the limes, because they were involved in the defence of the Empire. The hundreds of brand-new swords which have been found in Scandinavia and Germany, and partly also in Poland, must have been obtained as merchandise (Lønstrup 1988:96). It is unclear to what extent the Germanic warriors were equipped with swords at the beginning of our era. Behmer (1939:15) informs us that the Germans knew three types of swords: the one-edged hew-sword, the two-edged short Roman gladius and the long Roman two-edged sword, the so called La Tène III type, which was used by the Roman cavalry. This sword-type was the basis for the Germanic Migration Period sword (Behmer 1939:18). The one-edged sword was actually a big knife, a sax. The gladius is of Roman origin and was imitated by the Germans. Perhaps the puzzling word kesjam on the Bergakker scabbard mount may be explained by the assumption that the weapon designations for both swords and spears were confused. At the time the Bergakker inscription was made (early 5th c.), the word kesja may have denoted a Vennolum is a place in Norway, the findplace of the eponymous lance head. 36 51 certain sword-type; at a (much) later period the word got the meaning of ‘javelin’ (for another interpretation see the Checklist of Runic Inscriptions in The Netherlands). A (vulgar) Latin word for sword was CESA, the equivalent of Germanic *gaizaz (I guess the source was ultimately Celtic).

 An element such as Gesa- is found in the names of the Gaesatae and the Matronae Gaesahenae and Matronae Gesationum. A soldier of the Cohors I Vindelicorum was called Cassius Gesatus. According to Alföldy (1968:106) the name Gesatus is a cognomen, referring to the man's weapons. Probably, the Germans took over some special type of sword together with its foreign name. As to the tribe of the Gaesatae (recorded in 236 BC in the Alps), these people may have been Celts, so perhaps gaes- is a Celtic name for a Celtic La Tène sword. The lanceheads of the Illerup bog were of Scandinavian origin, made in Norway, according to Ilkjær, since an analysis of the iron points to iron ore from North Trondelag (personal communication). However, Roman know-how may have been wished for, a knowledge which may have been provided by Germanic weaponsmiths from among the foederati of the Rhine area. The obvious connection, then, is that wagnijo and niþijo learned their craft as weaponsmiths either in their homelands, or as mercenaries in the Roman army, where they also learned to sign their work. Where did they learn to do this in runes? In Norway? Unlikely. They probably learned this together with their craft. A runographic analysis shows a close resemblance between the runic graphs on the lanceheads (wagnijo) and the graphs on the shield handles (niþijo and laguþewa), which points to the same background of the runographers. Niþijo, as is mentioned above, had a workshop, where many of the Romaninspired items, found in the Illerup bog, were manufactured (Ilkjær 1996b:440f.). According to Ilkjær (1993) the lanceheads of the Vennolum-type36, to which the runic lanceheads belong, were widespread in Scandinavia. The runic spearhead from Øvre Stabu (2nd half of the 2nd c.) also belongs to the Vennolum type. Ilkjær states that only a few lanceheads from the Continent show some similarity, and that only one item from Poland is of the Vennolum type (personal communication).

THE FIRST RUNOGRAPHERS

Who could read and write runes in an almost illiterate society is subject of an often recurring debate. If one abandons the idea of a purely symbolical, magical or religious purpose of adding runes to objects, the answer is that at least the former mercenaries had learned to read and write, especially the officers. On the other hand there must have been literate people, more specifically craftsmen, among the foederati. The literate officers and soldiers must have constituted a small group. This would tie in very well with the observance that runic objects are sparse and emerge from widely separated places. Runic writing may have started as a soldiers’ and/or craftsmen's skill. This might explain the curious meaning of the word ‘rune': secret, something hidden from outsiders. The runic legends show very simple information, but it may be that the art of writing was sort of ‘secretive', the prerogative of a specific group only, and not necessarily linked to magic or religion. The application of writing, especially on precious objects points to special artisans. Signing one's name marks the pride of the author, who knows an extraordinary skill. He stands out in society because of his knowledge, and Syrett (1994:141) proposes to view swarta and similar instances, such as laguþewa as West Germanic strong 37 nouns with loss of final *-z. Here one apparently felt inclined to read the later Scandinavian h or A rune, and even a ‘repaired’ n rune has 38 been suggested (see Krause 1996:191, with ref.). 52 therefore obtains a special status. Naturally, he would be very reluctant to pass this knowledge on to others, which would make it more common. Perhaps this also (partly) explains the extreme rarity of objects exhibiting runic writing, dating from the early ages.

 THE WEST GERMANIC HYPHOTHESIS

An indication for a West Germanic origin of runic writing is the presence of West Germanic name forms on some of the oldest artifacts: wagnijo and niþijo (see above), harja (cf. Peterson 1994:161), swarta37, hariso, alugod, leþro, lamo (cf. Syrett 1994:141ff.), and also laguþewa. These attestations are from circa 200 AD and somewhat later, found in bogs and graves in Jutland, on Funen and on Sealand. Stoklund (1994a:106) points to the remarkable fact that all inscriptions that show West Germanic forms or which have West Germanic parallels are on weapons that originate from the area around the Kattegat, Scandinavia or North Germany and which were deposited in the Illerup and Vimose bogs. Few would claim that a West Germanic speaking people lived in those areas around 200 AD. But individuals such as weaponsmiths and other craftsmen, descending from a West Germanic speaking area, may very well have been present there. Especially the names ending in - ijo seem to point to the region of the Ubii in the Rhineland, since this was a productive suffix in Ubian names (Weisgerber 1968:134f.). The problem of the a- and o- endings, present in the nominative forms of apparently masculine names in runic inscriptions found in Denmark, has long been the subject of discussion. Syrett (1994:151f.) concludes that the early evidence, e.g. up to c. 400, "clearly indicates that -o and -a could be used side by side to represent the masculine n- stem nom. sg., but in the later period, as exemplified (...) by the bracteates, -a predominates". Herewith the case has not yet been cleared. Perhaps the problem should be tackled from a different angle. An examination of the recorded names of Germanic soldiers in the Roman army shows that the endings -a and -o are quite frequent. It may very well be that names featuring these endings were introduced to the North by veterans and craftsmen, such as weaponsmiths. As has been argued above, wagnijo and niþijo may have originated from the Rhineland, from the tribes of the Vangiones and Nidensis. The owner of the Vimose comb (with runic inscription) may have been a member of the tribe of the Harii, a sub-tribe of the Lugii. The descent of the man who wrote harja on his comb, is supported by a runic inscription on the Skåäng stone in Sweden, reading harijaz leugaz, evidently pointing to both Harii and Lugii. The reading harijaz is based on the assumption that the 7th rune is the z, corresponding with the ‘Charnay’ rune £ representing z. Its ornamental form has as yet not been recognised as the rune for z in this Swedish rune-text38. harja reflects a West Gmc dialect, with loss of final -z in the nominative. Just as in wagnijo and holtijaz the elements ijo and ija may be interpreted as an indication of someone's descent, harja can be interpreted as referring to someone belonging to the tribe of the Harii. A more extended form is the spelling harijaz of the Skåäng stone.

 The runes fir?a on Illerup VI may refer to the tribe of the Firaesi (Schönfeld 1965:88). Furthermore, one may 39 speculate as to whether the name harkilaz of the Nydam sheath plate contains a scribal error; perhaps it should represent haukilaz, provided the third rune should be read as u, not r (its shape, however, is that of an 'open’ r rune: _ ). If so, it could be interpreted as a reference to the Chauci. Besides, ON hark- ‘tumult’ is difficult to explain as a name-element. 53 suggested that the second part of this inscription leugaz was derived from the tribal name Lugii. Apparently Krause (1971:163) and Antonsen (1975:66) were not aware of the possibility of finding a tribal name here. The name Lugii appears to be related to Go *Lugj©s (Much 1959:378) and Go. liugan 'to marry', actually 'to swear an oath'. The root *leugh-, *lugh- ‘oath’ is only attested in Celtic and Germanic (Schwarz 1967:30). The Lugii, according to Much (1959:378), were a group of tribes, probably unified by an oath. The Harii lived in North Poland, not far from the Baltic. The comb may well have originated in that area, because of its find-context, which, according to Ilkjær (1996a:68), consisted of a combination of certain Polish fire-equipment "Indslag af pyrit og evt. polske ildstål", buckles with a forked thorn, and combs consisting of two layers, such as is the case with the harja comb (cf. the map in Ilkjær 1993:377 and further on the text on pp. 376-378). 7. Conclusions The Skåäng inscription supports the interpretations of wagnijo, niþijo and harja, as being appellativa referring to certain tribes, and not just personal names. According to Bang (1906:- 48f., note 419), Germanic PNs are often derived from tribal names. Other instances are the Hitsum (Friesland) bracteate (approximately around 500 AD), with the legend fozo, a PN, which may have been derived from the tribal name of the F©si (cf. IK, nr. 76, and the Checklist of Bracteates with Runes in the Catalogue), and the Szabadbattyán brooch, with the legend marings (see nr. 36 in the Checklist of Early Danish and Gothic inscriptions).

As to tribal names (attested in the Roman period) on Scandinavian stones, we have the forms haukoþuz (Vånga), hakuþo (Noleby). It may be useful to investigate once again the possibility, whether here the Chauci are referred to. Further there is ekaljamarkiz baij?z (Kårstad), perhaps pointing to the Bavarians? swabaharjaz (Rö) may refer to the Suebi, living on the right bank of the Rhine, iuþingaz (Reistad) to the Iuthungi (South Germany, north of the Danube), saligastiz (Berga) perhaps to the Salii (near the lower Rhine). Birkhan (1970:170, note 243) suggests the patronymic wagigaz on the Rosseland stone may contain the PN Vangio39. If these assumptions are correct, the inscriptions on the above mentioned stones may be dated rather early, on historical grounds, to between 200 and 500 AD. If wagnijo is exactly to be pronounced as Vangio, one has to accept the fact that the sequences of -gn- and -ng- both represent the sound [h]. In Roman ears the Germanic cluster gn may have sounded like ng. At any rate, the spelling of the tribal name Vangiones is in accordance with Latin practice. The same applies to the Roman spelling of the folk name Nidenses. Since the Romans did not know the graph þ, they most likely would write a d between vowels. Therefore, Nith- may be rendered Nid- in Roman orthography. Cf. also the cognomen Sinnio, a Germanic member of the corpore custos Drusinianus (Bellen 1981:73ff., note 40 105; and Weisgerber 1968:135, and 393f.). It may be that Sinnio shows West Gmc consonant-gemination, but on the other hand it might just reflect the name of the Roman gens Sinnius. 54 At some time in runic history there existed a rune W_to represent the sound [h], but it is not used to represent the sequence gn in wagnijo. Moreover, the carver applied W to render w: hence the (i)ng rune W may not yet have been present in the runic alphabet of around 200 AD. ¨

Masculine names ending in -io, n- and jan- stems, were especially frequent in the region of the Ubii, who were neighbours to the Vangiones. The names ending in -io reflect Germanic morphology representing the Latin ending -ius. The suffix -inius was reflected by Germanic - inio- (Weisgerber 1968:135, 392ff. and Weisgerber 1966/67:207). Weisgerber mentions the fact that within the n- stems of all IE languages we also find the on- type, which occurs in specific cases such as ion-, a type that is often met with in personal (Germanic) names (Weisgerber 1968:392). "Das Naheliegen von -inius bestätigt auch für das Ubiergebiet die Geläufigkeit der germanischen Personennamenbildung gemäß der n- Flexion. Mit dieser ist im ganzen germanisch-römischen Grenzraum zu rechnen. Die angeführte Reihe Primio usw. ist herausgehoben aus einer Fülle von Parallelbeispielen: Acceptio, Aprilio, Augustio, Faustio, Firmio, Florio, Hilario, Longio, Paternio usw." (Weisgerber 1968:394). In fact, in this way the question of the problematic ending -ijo in masculine PNs may be solved40. The awkward ending -a of laguþewa (cf. Syrett 1994:44f.) can be solved by accepting the fact that the name may indeed be West Germanic. Syrett states that even weak masc. forms such as swarta may be taken as West Germanic strong nouns, the "precursor of ON Svartr" (Syrett 1994:45). There is no need to postulate the presence of a runic koiné, such is suggested by e.g. Makaev (1996:63). He stated that: "Therefore the runic material, [...] provides important and elegant, albeit indirect, support for our hypothesis on the West Germanic-Scandinavian dialectal base of the runic koiné". One may simply change the term ‘runic koiné’ for ‘West Germanic origin of runic writing'. I cannot yet estimate the implications of the fact that the frequent occurrence of runic leub (and leubo, leuba, leubwini, lbi, leob, liub) in 6th century Germany may be connected with the many Leubo's in the area of the Ubii in the Roman period (Weisgerber 1968:150f., 167, 374f.). The name is also found among the Tungri and along the Lower Rhine. A runic attestation of the name is found in Västergötland, Sweden, on the SKÄRKIND stone: skiþaleubaz. This may refer to a Rhenish merchant of skins (containing the element ski(n)þa- ‘skin'). Another example is liubu (OPEDAL), but this may be no PN, but an adjective, or a verbform.

To sum up:

In view of the presence of (1) West Germanic name forms on the oldest runic attestations, and (2) the provenance of some of these objects, in combination with (3) the origin of the weaponsmiths wagnijo and niþijo, one may conclude that runic knowledge was first known on the Continent. (4) The inscriptions harja on the Vimose-comb and harijaz leugaz on the Swedish Skåäng stone confirm a connection between the North and the continental tribes of the Harii and the Lugii. (5) The presence of certain elite-weapons and -equipment in the Danish bogs is indicative of a network of contacts between elites from Scandinavia and the 55 Continent, and especially with provincial-Roman regions. The use of runes is closely linked to these relations. During the second century runic writing must have spread to the North. This is demonstrated by the runic brooches of Sealand, Jutland and Skåne, which were local products. The inscribed Vennolum-type lanceheads, including the lanceheads from Øvre Stabu and Gotland point to the possible presence of runic knowledge in Norway and Sweden, presumably taken there by Rhenish smiths. The weapon-trade between the Rhineland and the North may serve as evidence for close connections. I suggest the runic script was first developed in Romanized regions along the Rhine.


SOURCE: Runes around the North Sea and on the Continent AD 150-700, by J. H. Looijenga, 1997.