Myths and Literary Sources

The Old Norse word smiðr, meaning ‘smith’ or ‘crafter’, refers to the making of objects regardless of material, be it wood, bone, glass, or metals (Capelle, Reference Capelle, Blankenfeldt and Pesch2012: 17–18). In the mythological text known as the ‘Lay of Vǫlundr’, the smith boasts about his range of skills (Pettit, Reference Pettit2023: 341–64). The mythical Vǫlundr is a skilful ‘master of fine metals’, but also a man able to sharpen and temper his own sword. Vǫlundr has his equal in the dwarf smith Regin in the Poetic Edda and the late thirteenth-century Volsunga Saga. He describes himself as a master smiðr: ‘I knew how to work iron, as well as silver and gold, and from everything I could make something useful’ (Byock, Reference Byock1990: 57) (Figure 2); many such other artisans are mentioned in Old Norse texts dating from the ninth to the sixteenth century ad (Carstens, Reference Carstens, Blankenfeldt and Pesch2012; Marold, Reference Marold, Blankenfeldt and Pesch2012). Vǫlundr and Regin have their origins in Germanic legends, where they perhaps embodied a more widespread idea of multi-crafters in early medieval northern Europe (Byock, Reference Byock1990: 1–7; Hedeager, Reference Hedeager2011: 140–44, 177–81; Hardt, Reference Hardt, Blankenfeldt and Pesch2012). There are similar ideas in the Irish Mythological Cycle, e.g. in reference to the god Lugh as ‘a master of all arts’ (Old Irish Samildánach; Stokes, Reference Stokes1891: 77). In essence, a smiðr, in the original sense of the word, was a highly versatile artisan.

Figure 2. The dwarf smith Regin and his apprentice depicted on the doorway of the Hylestad stave church, Norway, dated to c. ad 1200. © Museum of Cultural History, University of Oslo, photograph by Ove Holst (CC BY-SA 4.0).

Several medieval written sources give descriptions of real-world metalworkers (for a discussion of ideal and real smiths of the Viking world, see Pedersen, Reference Pedersen, Lund and Melheim2009, Reference Pedersen2016). One of the most detailed is by Theophilus Presbyter, who wrote De diversis artibus in the first half of the twelfth century (Hawthorne & Smith, Reference Hawthorne and Smith1979). He gives us an insight into the variety of techniques employed in the production of precious items in gold, silver, copper, and tin, as well as iron. Theophilus also included chapters referring to the making of iron tools, where he describes forging and the hardening of steel through water-quenching (Hendrie, Reference Hendrie1847: 211–25). The level of detail and technical knowledge, as well as the designs of individual tools, suggest that Theophilus was himself a multi-crafter, even producing his own toolset. It also indicates that he was very much aware of the changing properties of iron through forging and of the techniques of iron soldering and brazing. Even decorative techniques, such as encrustation, that is used solely on iron objects and involves the application of precious metals, is discussed in Theophilus's treatise (Hendrie, Reference Hendrie1847: 381). The way Theophilus also included both ironworking and technically advanced processes aligns him with the smiðr Vǫlundr.

Graves, Hoards, and the Idea of the Multi-Crafter

In 1951, Jan Petersen counted as many as 375 Viking Age graves with metalworking tools in Norway alone (Petersen, Reference Petersen1951: 108). Most contained only a few such tools and may represent the toolsets used by the deceased (Bøckman, Reference Bøckman2007: 91) or symbolic attributes of high-ranking individuals (Ježek, Reference Ježek2015). There are also several Scandinavian examples of hoards containing toolsets and graves richly furnished with a large assemblage of both wood- and metalworking tools (Blindheim, Reference Blindheim1963; Müller-Wille, Reference Müller-Wille1977: 173–93; Arwidsson & Berg, Reference Arwidsson and Berg1999; Barndon & Olsen, Reference Barndon and Olsen2018). Blacksmithing tools are commonly found in combination with lightweight hammers, small tongs, small chisels, and sheet-metal shears, as well as small anvils for working softer metal (e.g. Blindheim, Reference Blindheim1963) (Figure 3). The combined toolsets have led to the suggestion that there was no clear distinction between the many types of metalworkers in the Nordic past (Petersen, Reference Petersen1951: 104; Müller-Wille, Reference Müller-Wille1977: 193; Jørgensen, Reference Jørgensen2012: 4; Guldberg, Reference Guldberg2014). Whether or not the burial was that of an actual (or ‘professional’) smith, the deposition reflects a status aligned with the mythological Vǫlundr. These graves and hoards may mirror the idea of the smiðr as a multi-crafter.

Figure 3. A metalworkers’ assemblage from a grave at Bygland, Norway. It contained a wide range of tools, from heavy sledgehammers (top left), small chisels, and punches (bottom left and right), to a long-handled iron pan for melting lead and tin (centre right) resting on a soapstone mould for casting ingots, as well as tools for working both ferrous and non-ferrous metals. © Museum of Cultural History, University of Oslo, photograph by Ove Holst. (CC BY-SA 4.0).

However, becoming a multi-crafter required the right conditions for the transfer of knowledge through social interaction. Undoubtedly, it must have involved long training. A skilled metalworker may well have begun training during childhood, and the understanding of materials and complex techniques must have taken years (Pesch, Reference Pesch, Blankenfeldt and Pesch2012: 41, with further references). While the multi-crafter was an ideal of the late first-millennium ad Scandinavia, probably only a limited number of people could have learnt to handle both iron and soft metals at a technologically advanced level. The development of skills must have been conditioned, requiring proper learning arenas and social mechanisms, such as those present in the royal courts mentioned in the sagas (Capelle, Reference Capelle, Blankenfeldt and Pesch2012: 17–18) and in Viking Age towns.

The Metalworkers in Towns, Marketplaces, and Farmsteads

From the 1930s onwards, an increased interest in Viking Age towns and early urbanism (see Hyenstrand, Reference Hyenstrand, Ambrosiani and Clarke1992: 39–40; Hilberg, Reference Hilberg2022: 38) is reflected in excavations that yielded an enormous body of material relating to production techniques, processes, crafts, and trade in Viking Age Scandinavia. While fine metalworkers and their products have received much attention (e.g. Armbruster, Reference Armbruster and Hines2004, Reference Armbruster, Blankenfeldt and Pesch2012; Feveile, Reference Feveile2006; Ambrosiani, Reference Ambrosiani2013; Pedersen, Reference Pedersen2016, Reference Pedersen, Glørstad and Loftsgarden2017; Orfanou et. al., Reference Orfanou, Birch, Sindbæk, Feveile, Barfod and Lesher2021), the working of iron (blacksmithing) and the way the blacksmiths’ roles and skill sets were integrated in this milieu have attracted less notice (see, however, Ottaway, Reference Ottaway, Ashby and Sindbæk2019). The extensive excavation campaigns of 1998–2003 at Kaupang did not reveal any in situ traces of ironworking or interrelations between different types of metalworking. It has nevertheless been suggested that non-ferrous and ferrous metalworking may have been carried out by different specialists at Kaupang (Pedersen, Reference Pedersen, Hansen, Ashby and Baug2015: 55), in view of the finds recovered at Ribe and Birka. Recent excavations at Ribe have uncovered traces of both ferrous and non-ferrous metalworking in the same workshop floor, dating to c. ad 810–830. The key activity though seems to have been the casting of soft metals—at least in one part of the building (Croix et. al., Reference Croix, Deckers, Feveile, Knudsen, Skytte Qvistgaard, Sindbæk, Wouters and Sindbæk2022: 174–75, 179).

Non-ferrous metalworking activity is somewhat easier to date typologically than pieces of iron slag. Metalworking tools and casting waste from soft metals may also be less difficult to link directly to processes and products. Further, iron tools, ingots, and slag/smelts of precious metals are more often recovered in metal-detecting, whereas iron slag may be overlooked (Gustafsson, Reference Gustafsson2013: 23). These factors affect the balance, availability, and compilation of key data for research into the presence and mechanics of multi-crafters in Viking Age markets and towns.

The Urban Scene and Viking Age Kaupang

Essential for what is today regarded as the ‘Viking world’ was a network of proto-urban nodes in north-western Europe connected by trade across the North Sea and the Baltic during the ninth and tenth centuries ad (Sindbæk, Reference Sindbæk2007). These centres were commonly founded and protected by powerful nobles and served as hubs for local and regional crafts and exchange (Sindbæk, Reference Sindbæk2007: 127). The remains of buildings, a vast number of artefacts, cultural deposits, and associated graves attest to long-distance trade and contact across the known world.

Kaupang, on the south-eastern coastline of present-day Norway, was one such centrally located nodal point. It is mentioned, around ad 890, as Skiringssal by the tradesman Ottar in a report to King Alfred of England (Skre, Reference Skre and Skre2007: 28–29). It was later named Kaupang, derived from the old Norwegian term kaupangr, meaning ‘market’ or ‘trading place’ (Brink, Reference Brink and Skre2007: 63).

The central part of Kaupang, the so-called ‘Black Earth’, consists of cultural deposits encompassing a 650 m-long belt along the western bank of the inlet of Kaupang (Figure 4, top). The size of Kaupang during the Viking Age is estimated to cover approximately 20,000 m², surrounded by a 34,000 m² zone with areas of more temporary trade and craft activity. Since 1956, around ten per cent of these areas has been excavated, including a harbour area (Pilø, Reference Pilø and Skre2007a: 130–31; Reference Pilø and Skre2007b: 152–54). Kaupang is interpreted as a permanent settlement from c. ad 800 to 930. Thereafter, trading and crafting appears to have continued only to a limited extent for another forty years before the site lost its function as a place for artisans and traders (Pilø, Reference Pilø and Skre2007d: 177–78).

Figure 4. Top: Kaupang and the site of Kaupangveien 224. Bottom: the site's main archaeological features, including a well-defined domestic building (‘house plot’), associated with the workshop and waste pit.

The main layout of Kaupang comprised small rectangular plots separated by ditches and fences, similar to most other urban Viking sites (Skre, Reference Skre, Brink and Price2008: 88–89, with references). It is estimated that some 90–100 plots existed within the central area along the bay, each plot large enough (approximately 40–80 m²) for one building and a small outdoor area (Pilø, Reference Pilø and Skre2007c: 193; Reference Pilø and Skre2007d: 178; Skre, Reference Skre, Brink and Price2008: 89). The excavation campaigns of 1998–2003 uncovered six plots, five of which contained house remains of a relatively uniform type (Pilø, Reference Pilø and Skre2007c).

Unn Pedersen has conducted extensive studies of the metalworking activity at Kaupang (Pedersen, Reference Pedersen, Lund and Melheim2009, Reference Pedersen, Hansen, Ashby and Baug2015, Reference Pedersen2016). While the excavations of 1998–2003 did not reveal any well-defined workshops with an intact furnace, evidence of fine metalworking was identified in two stratified contexts: a small part of a floor layer (plot 1B) and in house A302 (plot 3A). Because no iron slag was recovered in the floor layer of plot 1B, it was concluded that this may have been a specialized workshop for silver casting, as well as other soft metals. The finds and stratigraphy suggest a date to the second half of the ninth century for metalworking activity in plot 1B (post-ad 863) and pre-ad 840/850 for house A302 (Pedersen, Reference Pedersen2016: 182–84).

The Kaupang complex may originally also have contained about 1000 graves (Stylegar, Reference Stylegar and Skre2007: 75–78). Nearly 200 have been excavated, displaying diverse burial customs, as well as a close relationship to the town’s trading and production activities. The assemblage of artefacts of iron and soft metals from these burials is rich and varied, some showing excellent craftsmanship. Numerous graves surrounding Kaupang contained tools related to different types of metalworking and have been linked to activities in the town (Blindheim, Reference Blindheim, Blindheim, Heyerdahl-Larsen and Tollnes1981: 44).

The Kaupang Workshop

During the summer of 2015, a small rescue excavation of c. 160 m² in the south-western part of Kaupang (Figure 4, bottom) revealed, among other finds, a domestic building with roof-bearing posts, a curved wall ditch, and a hearth; its plan suggests a possible link to the Dublin-style houses of the Viking Age (Wallace, Reference Wallace1992: 100). The building was radiocarbon-dated to 1211 ± 29 bp, cal ad 702–891 (at 95.4 per cent confidence; Ua-53944, charred barley from a posthole) and 1218 ± 29 bp, cal ad 692–888 (at 95.4 per cent confidence; Ua-53945, charred barley, central hearth). Directly to the north-west of this building and at the same discovered plot, a well-preserved metal workshop with a furnace and an associated waste pit with metalworking debris was uncovered.

The workshop area was quite small and sub-rectangular, measuring approximately 2.6 × 2 m. The activity layers were well defined, suggesting that the workshop was enclosed by walls. An earlier well, lined with wooden planks, was situated underneath the workshop; it had been plugged before the workshop was built and was slightly truncated in the process. This process created a depression—which made it possible to construct a sunken floor in an easy and opportunistic way. Postholes suggest that timber posts supported a makeshift roof, and two juxtaposed postholes in the workshop's north-western corner may have been related to a possible entrance. A shallow, elongated waste pit lay directly outside this entrance, an expeditious place to dump waste from metalworking when cleaning the floor.

Inside the workshop, features include a 1.6 × 0.8 m dual-pit forge. The forge itself consisted of a shallow pit connected to a deeper pit, both lined with clay and located at the far end, opposite the entrance. The forge was dug into the sandy backfill of the well beneath the workshop. A modern disturbance had cut the southern part, truncating a small part of the forge (Figure 5).

Figure 5. The metal workshop, view towards north-northeast. The clay-lined dual-pit forge at the bottom of the image is only partly excavated, and sandy soot- and charcoal-rich floor layers lie directly north of the forge. To the left is the circular cut of the underlying well.

Next to, and around the forge, several charcoal-rich layers contained various amounts of metalworking debris. Analysis of soil sampled in a limited grid from these soot- and charcoal-rich floor layers (A2612/R2612 and A2713/R2713) showed the presence of hammerscales and slag spheroids from ironworking, in combination with crucible fragments (n = 61) and other forms of debris from handling soft metals (Jouttijärvi, Reference Jouttijärvi and McGraw2017: 18–32, 38–42, see also below). In contrast, two layers, interpreted as the fill of the workshop itself (A2425, A2426), contained fewer crucible fragments (n = 30), little slag, and much less charcoal. The largest assemblage of crucibles was collected from the forge itself, with 187 fragments. As for tools, only a small, well-used chisel was recovered.

The workshop has been dated by multiple methods. The well beneath the workshop, lined with vertical planks of oak of variable size and thickness, displayed signs of reuse and some had holes filled with wooden pegs. The planks were dated by dendrochronology, with felling dates in c. ad 800–805 and during the winter of ad 823/824 (Daly, Reference Daly and McGraw2016) (Figure 6), which provide a solid terminus post quem for the metalworking activity. A charred fragment of hazelnut from the larger dual-pit forge was radiocarbon-dated to 1046 ± 31 bp, cal ad 895–1038 (at 95.4 per cent confidence; Ua-53943) and an uncharred hazelnut fragment in a shallow pit at the entrance was dated to 1140 ± 30 bp, cal ad 774–992 (at 95.4 per cent confidence; Ua-53942). The uncharred hazelnut lay next to a large brass ingot (cf. Figure 4) and was preserved due to the antibacterial qualities of copper alloy. The ingot is similar in size and shape as examples from a hoard in Hedeby, a type dated to the eighth and early ninth century ad (Sindbæk, Reference Sindbæk2003: 57–58). The same pit also contained an oblate spheroid weight with simple dot decoration of a type more common after c. ad 860/870 (Gustin, Reference Gustin2004: 314). Fragments of five different beads, of which two belong to Johan Callmer's (Reference Callmer1977) group Av, type A340T (ad 835/840–875/890), and one to group Bc, type B545 (ad 830/840–860/870), were found in the workshop; although only bead B545 was collected from an in situ layer. The rest were recovered from the backfill layers, and their typological dating and context suggest that they were redeposited.

Figure 6. Section through the workshop and underlying well.

The combination of these dates, together with stratigraphic data, suggests that the workshop was not established before ad 824. While the timeframe between the oak felling and the reuse of the timber in the well is difficult to estimate, the radiocarbon data support a dating to the latter part of the Viking Age: the latest ¹⁴C date of cal ad 895–1038 on a fragment of hazelnut shell retrieved from a sealed context, with additional fragments found in the same micro-context, lends further reliable support.

The recovery at Kaupang of typologically datable finds during the 1998–2003 investigations of non-stratified deposits suggests that occupation of the settlement ended in c. ad 960–980, although later cultivation has eradicated stratified deposits later than ad 840/850. The workshop thus seems to have been in use in the later period of occupation at Viking Kaupang, possibly during a period previously thought to have been destroyed by subsequent farming (cf. Pilø, Reference Pilø and Skre2007d: 177–78). Its sunken floor saved it from eradication by later agricultural activity.

Ironwork

Iron slag, weighing 3.5 kg, was retrieved from the workshop and associated waste pit (Figure 7.4). Approximately 2.6 kg was collected from the elongated waste pit and 0.7 kg from various layers within the workshop itself. The waste pit contained the largest pieces of slag, two of which resembled small blooms. No clear stratification was observed within the pit, and fragments of crucibles, in addition to a few pieces of copper alloy and a piece of gold, were uncovered in the same layer as the iron slag.

Figure 7. A chisel (1), a possible bell-hammer (2), an iron spring fastener for oval brooches (3), and slag from working iron (4) and fine metal (5). © Museum of Cultural History, University of Oslo, photograph by Vegard Vike (CC BY-SA 4.0).

Six pieces of ferrous slag were metallurgically analysed using scanning electron microscopy with energy dispersive X-ray spectroscopy (SEM/EDX) to determine their composition, three from the workshop and three from the waste pit (Jouttijärvi, Reference Jouttijärvi and McGraw2017: 42, 127–34). The workshop also contained items such as rivets, nails, cramp irons, a bolt, and a fishhook, all of iron and possibly made in the workshop.

Microscopic finds of hammerscales and slag spheroids suggest that both refining of raw iron (bloom) and working of refined iron took place in the workshop. Moreover, the analysis of selected pieces of iron slag from the waste pit and the workshop displayed similarities in composition and structure. With one exception, the analysis suggested that the refining/working of blooms probably came from the same iron extraction event or the same source of bog iron. One piece of slag contained a higher amount of manganese oxides, suggesting extraction from another bog source (Jouttijärvi, Reference Jouttijärvi and McGraw2017: 42). The results indicate one or two episodes of refining iron and the similarities of slag composition provide insights into the relation between the debris in the pit to the activity inside the workshop. This gives us a rare glimpse into the upkeep and maintenance of the workshop during its lifespan.

Iron refinement indicates that the raw iron may have been brought here directly from an extraction site. More recent sources of traditional blacksmithing highlight the importance of refining as integral to the production of good quality iron (e.g. Evenstad, Reference Evenstad1790: 437–40) and of the role of the blacksmith as a quality assessor of good workable iron.

Soft Metals and Technical Work

Soft metals were recovered in various amounts in the workshop and the adjoining waste pit, from the same contexts as the ironworking debris (Figure 8). The evidence for working soft metals consisted of nearly 500 fragments of crucibles (0.42 kg), approximately 300 fragments of burnt clay (1.2 kg), pieces of clay moulds, two offcuts of silver and one of gold, as well as melting debris from copper alloys and lead (Figure 9).

Figure 8. Selected metalworking finds from the workshop. (1) a 190 mm-long brass ingot; (2) offcut of gold with spirals in false filigree; (3) offcuts of brass; (4) a spheroid weight with simple dot decoration; (5) offcut of silver. © Museum of Cultural History, University of Oslo, photograph by Vegard Vike. (CC BY-SA 4.0)

Figure 9. Poorly preserved clay moulds with imprints (top left), small crucible fragments (right), and fragments of charred hazelnut shells (bottom left), all from the workshop. © Museum of Cultural History, University of Oslo, photograph by Vegard Vike (CC BY-SA 4.0).

The metallurgical analysis of thirty-one crucible fragments, of which sixteen were from the workshop itself and fifteen from the adjoining waste pit, revealed that three different types of temper had been used in producing the crucibles. One group was exclusively tempered with quartz, one with feldspar, and the third group with a combination of both. Soft metals like gold and silver had been worked, but far less frequently than brass, bronze, and lead-bronze alloys (Jouttijärvi, Reference Jouttijärvi and McGraw2017: 5). The use and work of such non-ferrous alloys are in line with previous analysis of metalworking waste recovered at Kaupang in 1998–2003 (Pedersen, Reference Pedersen2016: 189–94, fig. 5.1).

Four fragments of crucibles used in cupellation were identified, supporting the case for small-scale refining in the workshop, perhaps of impure silver (see Bayley & Eckstein, Reference Bayley, Eckstein, Sinclair, Slater and Gowlett1995). The process of cupellation involves the removal of impurities by adding lead to liquefied metal. A spongious material (e.g. charcoal or bone) can be added. The oxidizing lead acts as a binding agent for impurities, thus encouraging the purification of silver for further recycling and/or quality assessment. The cupel fragments had a mixture of quartz and feldspar as a temper, with distinct layers of lead oxides clearly visible on X-ray. Crucibles densely tempered with quartz have previously been identified at tenth-century ad Fyrkat, where they had been used in silver refining (Lønborg, Reference Lønborg1998: 14–15).

A small group of crucible fragments, made of a thicker material, stood out from the assemblage. While also made of fired clay tempered with quartz and feldspar, their cross-section showed a layered encrustation, one dark grey and one with bright red/purplish oxidation. The metallurgical analysis of one piece indicated markedly higher levels of evenly dispersed silver in the dark grey layer, as well as silver sulphides. This combination indicates that niello was produced (Jouttijärvi, Reference Jouttijärvi and McGraw2017: 10–11), and may suggest that technically advanced methods of decoration took place in the workshop.

Micromorphological studies of the charcoal-mixed layers in the forge found evidence of both ferrous and non-ferrous metalworking in the form of trace elements embedded in the same micro-contexts: siliceous glassy slag and trace amounts of zinc were found, possibly deriving from marine plant material, suggesting either the use of fluxes in soldering or a source of fuel in the workshop (Macphail & Linderholm, Reference Macphail, Linderholm and McGraw2016: 2, 8–11).

Most of the clay casting moulds were poorly preserved, as were those from earlier excavations (Pedersen, Reference Pedersen2016: 181–83), and very few reveal any indications as to what was cast in the workshop. The imprint of part of a pin in one of the mould fragments (Figure 9) hints at the production of simple stickpins—a common type of jewellery in the western Viking world (Graham-Campbell, Reference Graham-Campbell and Skre2007). An iron spring fastener (Figure 7.3), typical of oval brooches, suggests that the production or repair of such jewellery also took place.

Three fragments of metallurgical clay packages with faint traces of textile or organic imprints show that brazing of iron may also have been carried out. The imprints are from wrapping iron parts with strips of copper alloy, enclosing them in clay packages, and heating them in the forge in order to braze the iron parts together (see Söderberg, Reference Söderberg2004). Such fragments have been identified in earlier excavations at Kaupang as possibly representing the production of padlocks (Pedersen, Reference Pedersen2016: 135–40). Large-scale production of padlocks and small bells has been suggested for Birka (Gustafsson, Reference Gustafsson2005), and spherical weights were produced in similar clay packages (Söderberg, Reference Söderberg1996), indicating that similar techniques for dissimilar objects were employed in the same milieu.

The 190 mm-long brass ingot mentioned earlier (Figure 8.1) is similar to ingots from Hedeby (parallels to ingot hoards from Myrvälde and Kamänget on Gotland in Sweden are also worth noting, though the Gotlandic types are much larger). Clay moulds from casting Hedeby-type ingots have been found at Ribe (Sindbæk, Reference Sindbæk2003). Recent lead isotope analysis of ingots from the Hedeby hoard indicate that they must have been produced at a high level of standardization, possibly from Balkan copper ores (Merkel, Reference Merkel2018), suggesting that specific workshops had direct access to raw materials, through organized production, and that metalworkers contributed to a market-driven economy. The Kaupang workshop may have been part of this network.

The smithing debris in our Kaupang workshop attests to varied techniques, processes, metals, and alloys, reflecting shared customs and contacts, as well as ways of doing between different Viking workshops. In turn, this may mirror learning traditions and diversity, reflecting the metalworkers’ individual environs.

Multi-Crafters or Interdisciplinary Collaborators

Initially we addressed the idea of the smiðr, the multi-crafter of the Viking world, and argued that Viking graves and hoards could reflect such an idea. The excavation of the Kaupang workshop indicates that highly technical and diverse metalworking processes took place within the same workshop, raising the question of multi-crafting.

Mixed contexts with traces of both ferrous and non-ferrous metalworking are not unique for Kaupang. Indications of similar combined workshops have been identified in major Viking trading centres such as Birka (Ambrosiani, Reference Ambrosiani2013: 223), Dublin (Wallace, Reference Wallace2016: 311), Hedeby (Schietzel et al., Reference Schietzel, Bau, Hammon and Schüller2014), Ribe (Brinch-Madsen, Reference Brinch Madsen and Bencard1984: 79–90, cf. Reference Brinch Madsen, Bencard, Brinch Madsen and Rasmussen2004: 206–10), and York (Ottaway, Reference Ottaway1992: 719), and examples are known from outside urban towns (Gustafsson, Reference Gustafsson2011). A workshop in the late Viking Age settlement of Viborg Søndersø in Jutland (Denmark) offers the most detailed insight into the combination of such activities. Studies of its well-preserved microtopography showed evidence of varied use (Jouttijärvi et al., Reference Jouttijärvi, Thomesemn and Moltsen2005; Jouttijärvi, Reference Jouttijärvi2014), including iron smithing, casting of silver, bronze, and lead, as well as comb-making, i.e. evidence of multiple crafts. Pieces of cupellation crucibles were also found, as at our Kaupang workshop, although in small amounts, suggesting that silver refining may also have taken place. The range of products made in the same workshop during the same season may suggest that the Viborg workshop was used by more than one craftsperson. It is, however, impossible to determine how many were present in any one working season, despite the excellent state of preservation and documentation.

The ability to exercise multiple crafts is a question of practice and schooling. It is, however, difficult to estimate the degree of practice and training the metalworkers needed (Pesch, Reference Pesch, Blankenfeldt and Pesch2012: 41), but it is unlikely that an average artisan could have mastered all specialized forms of metalworking in equal measure. Therefore, the production of certain complex artefacts may have taken place within a larger operational network of specialists, driven by the development of urbanism and/or market-based economies (Croix et al., Reference Croix, Neiß and Sindbæk2019: 358; see also Pedersen, Reference Pedersen, Pedersen and Sindbæk2020). Furthermore, Viking Age metalworking has been considered a social and cooperative activity, with workshops potentially shared by several people (Callmer, Reference Callmer, Hårdh and Larsson2002; Sindbæk, Reference Sindbæk and Sigmundsson2009; Pedersen, Reference Pedersen2016: 34). Consequently, it can be argued that the mixed metalworking debris in the workshop at Kaupang reflects the work of multiple artisans, potentially with different skills and specialisms.

However, there are also strong arguments in favour of the concept of the multi-smiðr in the real world, and that the possession of multiple skills was seen as an ideal that allowed and encouraged multi-crafting. Traces of non-ferrous metalworking have been documented in rural smithies in Norway and Sweden, sometimes far away from urban contexts. The evidence for non-ferrous production at these sites is, however, often limited, suggesting an absence of separate traditions of metalworking (Narmo, Reference Narmo1997: 152; Rødsrud & Jouttijärvi, Reference Rødsrud, Jouttijärvi, Rødsrud and Mjærum2020; Sahlen, Reference Sahlen, Karlsson and Magnusson2020). When required, rural smiths also handled soft metals as well as refined iron blooms. The Kaupang workshop, with its mixed metalcrafts, was therefore by no means unique in the Viking world.

The vast production of composite objects, i.e. products of combined metals and techniques, also suggests an absence of a strongly divided metalworking tradition during the Viking Age. Composite items are good indicators of collaborative work or products reflecting the presence of a multi-crafter. With an iron core, swords, spearheads, and prized tools of iron were often decorated with precious metals using various techniques. These objects would have required combined skill sets and multi-crafting knowledge. The same can be argued for copper-coated iron weights and the soldering/brazing of padlocks, which require complex processes combining ferrous and non-ferrous metalworking techniques (Söderberg, Reference Söderberg2004; Gustafsson, Reference Gustafsson2005; Ambrosiani, Reference Ambrosiani2013: 223; Pedersen, Reference Pedersen2016: 135–40). Many artisans were even proficient in both technique and art, applying the animal and vegetal styles of ‘the Viking way’ to various materials, such as iron, soft metals, wood, and stone (e.g. Wilson, Reference Wilson, Brink and Price2008), thereby integrating forms of craftsmanship into one common tradition.

The archaeological record suggests that the skilled artisans and other metalworkers at urban sites, such as Kaupang, had access to collective spaces, where they could share their trade, knowledge, and experience. Such arenas may have resulted in, possibly even enhanced, the social and artistic mobility of knowledge and experience needed to be a true multi-crafter, which, in its idealized form, is reflected in the myths and ideas of the smiðr graves. The evidence from the Kaupang workshop and elsewhere that attests to the multiple skills and experience needed to produce certain goods, make it highly likely that multi-crafters existed in a Viking world, in which there was no strict separation between those who worked with iron and those who worked with soft metals.

Concluding Remarks

The data collected from a metal workshop excavated at Kaupang in 2015 indicate that artisans worked with both ferrous and non-ferrous materials, handling advanced technical processes in addition to refining iron. The context suggests cross-crafting between technical and mechanical disciplines. The degree of material knowledge is significant, in terms of the economy (i.e. silver assaying and iron refinement) as well as in the recycling of metals for casting and blacksmithing. The multi-crafters identified at Kaupang and in other workshops display versatility, mirroring the body of ideas represented by the assemblages of tools in contemporary Viking graves. Traditionally, there has been a tendency to study specific crafting traditions in isolation, but more nuanced, interactive, and cross-disciplinary approaches to the study of crafts are needed (Pedersen, Reference Pedersen, Pedersen and Sindbæk2020: 401–02). In the Viking Age, market-based crafts emerge in an economically co-dependant society. The graves and hoards can be seen as social expressions of, and responses to, these markets and networks, expressing value, power, and materiality. These markets gave artisans a direct link to both materials and consumers, creating the basis for crafting mobility through ‘learning by doing’.

Of course, not all artisans possessed the same degree of control, knowledge, and skill with metals and materials. Many Viking Age metalworkers in rural areas may not have had access to the knowledge or experience that would allow them to undertake complex metalworking. Others must have been professionalized, as well as highly specialized, producing the most valuable, high-status metalwork. These metalsmiths possessed a technological expertise in line with that conveyed by Theophilus in the twelfth century and may have had a status on a par with Vǫlundr, which could have originated at the same time as the increased demand for high-status objects at the onset of the Nordic Late Iron Age (c. ad 550–800). Finds from urban contexts, such as those from the Kaupang workshop, with advanced and mixed metalcrafts under the same roof, may reflect the work and presence of multi-crafting. These urban nodes were places that allowed multi-skilled artisans to develop and prosper, in essence places that enabled the smiðr to become something close to the ideal of the myths and idols of the Norse mythological world.