Objectives and methods

The core objectives of this study are twofold. First, to empirically interrogate the engagements anthropomorphic objects elicited and the practices in which they may have been caught up. By looking to traces of attachment, handling or manipulation we ask, in what ways did these iconic Viking Age objects intra-act with the world? The article thus examines traces of engagement, use and breakage through macro- and microscopic analytic approaches from a theoretically engaged perspective.

The second objective, explicitly relating to Viking Age discourse but with transferable value, is to trouble the way in which scholarship has enforced gender and other categorisation in the interpretation of these artefacts. For example, so-called ‘valkyrie pendants’—anthropomorphic figures conventionally interpreted as mythical female beings who guided fallen warriors to the deathrealm Valhalla—have been assumed to be worn by women (see discussion below). We aim to take a step back from the familiar, comfortable categories Viking Age scholarship often reverts to (binary-gendered worlds of housewives, kings and warriors; cf. Arwill-Nordblahd Reference Arwill-Nordbladh2013a; Marshall et al. Reference Marshall, Thompson, Wilson and Eriksenin press) and think anew about these miniature metal bodies. What happens if we stop forcing artefacts into our categories, and instead let them show us how they engaged with tools, hands and spaces—how they changed over time, demanded attention, were broken and remade?

Towards the fulfilment of these objectives, we employ two methods: microwear analysis and RTI (see online supplementary materials (OSM1–3) for additional detail). Microwear analysis is the macro- and microscopic study of traces that develop on the surface of objects during their manufacture, handling, use and post-depositional taphonomy. Primarily this analysis is applied to the study of tools and ornaments (e.g. van Gijn Reference van Gijn2010; Crellin Reference Crellin2018), and to a limited extent to figural objects (e.g. Bider et al. Reference Bider, Hamon, Bellot-Gurlet, Beyries, Pradeau, Vautier and Voldoirem2014). Objects were subjected to low power microscopic wear analysis (up to 50× magnification) using a stereomicroscope (Nikon Nissho Optical) and micrographs were taken using a Dino-Lite Universal Digital Microscope and DinoCapture 2.0 software. Recorded wear traces include linear features (striations, troughs), surface modification (percussive traces, abraded areas) and deformation of perforation rim (facets, edge rounding).

RTI captures surface topography through photography and computational processing, recording three-dimensional reflectance properties of object surfaces. Objects were photographed with a light source positioned at different angles, and the images processed to reveal surface details at enhanced visibility. Application of raking light and alternative surface renderings can help determine details in decoration, manufacturing processes and post-manufacture interaction (e.g. Jones & Díaz-Guardamino Reference Jones and Díaz-Guardamino2019; Min et al. Reference Min2021).

Previous interpretations

While representational approaches provide invaluable insight into the symbolism and ritual power of anthropomorphic objects (e.g. Arwill-Nordbladh Reference Arwill-Nordbladh, Back Danielsson and Thedéen2012; Helmbrecht Reference Helmbrecht, Hårdh and Larsson2013; Price Reference Price2019), we focus here on their more-than-representational qualities, specifically the material traces of intra-actions. A preoccupation with visual symbolism has, we argue, led to three core assumptions about these iconic objects. First, each of the 10 objects studied here has, in scholarship or curatorial cataloguing, been connected with mythological beings from later written sources (e.g. Thor, Freya, Mimir, valkyries) or with aristocratic housewives and warriors (see OSM4). Objects O5–O10 are recorded in the museum catalogue as depicting ‘valkyries’, yet variation in iconography and material raises questions over whether the objects constitute a single motif. Homogenising typologies may bring implicit presuppositions about homogeneity of use.

Second, and consequently, each object has been assumed to have a symbolic-magical purpose or to constitute an ‘amulet’ (Table 1). Two objects (O1 & O2) have been interpreted as ritual paraphernalia for an Old Norse völva, a female ritual specialist (Price Reference Price2019: 103–104). Third, all objects except O3 are characterised as body adornments (e.g. Plochov Reference Plochov and Fransson2007) and are assumed to have been worn by women. For example, in their otherwise innovative article on digital reconstruction of figurine moulds, Deckers and colleagues (Reference Deckers, Croix and Sindbæk2021: 52–53) assume rather than argue that ‘valkyrie’ pendants were worn by “women … on ceremonial occasions … emphasis[ing] their role and status”. This is despite the fact that only two of the six ‘valkyries’ discussed here were found in burials that potentially also contained the skeletal remains of female individuals (which, in itself, is not conclusive evidence for gender-specific wear in life). Moreover, if the figures do depict valkyries who guided warriors to Valhalla, would we not expect that male warriors could also wear them? Assumptions of body adornment as inherently ‘feminine’ may be bleeding into these interpretations (O’Sullivan Reference O’Sullivan2015).

Scholarship has generally been preoccupied with gendering late-prehistoric Scandinavian anthropomorphic art through a representational framing, despite the lack of primary sexual characteristics on most figures (cf. Arwill-Nordbladh Reference Arwill-Nordbladh2013a). This has generated interpretative guidelines such as, for example: ‘women often had hair that was long or in knots’; ‘men usually wore trousers or had delineated legs’; ‘couples are usually male-female’ (Mannering Reference Mannering2017). While important as foundational research, we do see three issues with such categorisations. First, despite critical interventions (Back Danielsson Reference Back Danielsson2007; Arwill-Nordbladh Reference Arwill-Nordbladh2013a; Amundsen & Moen Reference Amundsen and Moen2023), a binary-gender system is conventionally assumed, not critically examined. Second, to make these systems work, scholarship has acknowledged, then excluded, a significant number of ‘ungendered’ objects (e.g. Mannering Reference Mannering2017: 26–27). Finally, the warrants of ‘often’ and ‘usually’ tend to disappear in subsequent consideration. Consequently, we are left with somewhat self-affirming iconographic traits that gloss over complexity and contradiction. When objects that defy easy sex/gender classifications are excluded from analysis, representational frameworks can tie us into particular kinds of epistemes while suppressing others. This article thus seeks to ask other research questions to generate other types of knowledge.

Manufacture and multistaged decoration

The figures analysed were created through a combination of techniques including casting, polishing, engraving and stamping. Moulds for anthropomorphic objects are known from key Viking Age sites such as Birka, Sweden, and Ribe, Denmark (Deckers et al. Reference Deckers, Croix and Sindbæk2021: 40). Moulds were ceramic and generally single use, being broken to extract the figure. Consequently, body-makers could produce related, but rarely identical, objects. At the casting stage, body-makers decided which traits to accentuate or suppress; among the 10 objects studied, facial features are generally less detailed, while garments are rendered in a more intricate manner. Possibly, facial features were deemed less important than the costume for the intended purpose of the object.

Eight of the objects analysed were cast in silver, a highly valued material for the Vikings. Acquisition of silver was a driving force in Viking raids and expansions; the metal was sourced mainly from Arabic dirhams that were melted to produce jewellery and other objects (Kershaw & Merkel Reference Kershaw and Merkel2022). The capacity to recycle and remake metals is integral to the biographies of these miniature bodies; gleaming objects from faraway lands ultimately stabilised into bodily forms. The process of metal body-making required complex assemblages (clay, fire, hammers, chisels) and involved the entire sensory apparatus, assessing colours, sounds and qualities of the smelt at different stages (Kuijpers Reference Kuijpers, Stig Sørensen and Rebay-Salisbury2008). Most of the objects studied here were presumably made by different hands, and different moulds, in order to rearticulate the human body in various forms and substances. Yet, some seem to share histories of manufacture—perhaps making them ‘kin’ (Eriksen Reference Eriksen2022).

The process of making did not finish with casting but entailed further steps, principally the use of abrasion (with progressively finer materials) to create smooth, shiny surfaces. Variation is apparent in the degree of polish between the obverse (front) and reverse (back) sides of the objects, and in the way in which grooves were enhanced. Figures O2, O6, O8 and O9 saw detailed polishing of both obverse and reverse sides, while some irregularities from casting are still visible on the reverse of O4 and O5, and O7 saw polish only on the obverse side. While such variations may be interpreted as varied levels of skill, they may also be linked to human-object intra-actions. Did polishing choices connect with how figures were intended to be displayed, and whether the reverse was expected to be visible? Or were traces of earlier stages of manufacture deliberately left as evidence for the processes of making associated with the figure?

The Rällinge figurine (O3) stands out in terms of making (Figure 3). Following casting, its surface was smoothed using an abrasive material, and chasing with a hammer and a punch was then used to define certain features, including grooves below the eyes and spirals on the back. Occasionally, the groove interiors were subsequently smoothed over and regularised using abrasion. The spiral decorations are not consistent in their execution, however. On the right shoulder, a double spiral is seemingly in the process of making, with one of the lines not fully chased. The opposite shoulder has only one spiral, while two curved lines rather than full spirals mark the lower back. The variable execution in these details may suggest that people engaged with the figure by adding elements at different times. The process of making, therefore, may have been extended, opening tantalising possibilities for how to understand the intra-actions among human and metal bodies. Rather than a linear sequence of actions (completion of manufacture, use), engagement with human maker(s) may have happened through multiple events.

Figure 3. Microwear traces on O3, the Rällinge figurine: A) manufacturing striations; B) groove below eye; C) incomplete spiral groove on the upper back; D) single curved line on the lower back (figure by C. Tsoraki; Rällinge photograph by O. Myrin, Swedish Historical Museums, CC BY 4.0).

Attachment, circulation, wrapping

For most of the objects examined here, scholarship has assumed that they were displayed as pendants or ‘amulets’ (e.g. Helmbrecht Reference Helmbrecht, Hårdh and Larsson2013; Deckers et al. Reference Deckers, Croix and Sindbæk2021). Indeed, nine out of 10 objects display some form of suspension device—perforation through the object or loops on the reverse or on top of the object—indicating the capacity to be assembled with other objects and materials (conventionally assumed to be strings of beads, necklaces or other jewellery). They were thus (re)configurable and manipulable, able to enter in and out of different kinds of relations and collections. Only the Rällinge figurine was made without an obvious device for attachment.

To investigate potential attachments, we examined the attachment loops and perforations for microwear traces. Our results demonstrate a high degree of diversity in attachment rather than a homogenising interpretive function as ‘pendants’ (Figure 4). Three objects (O5, O7 & O10) show no traces of suspension from their attachment loops, indicating limited or no use as ‘pendants’.

Figure 4. Micrographs showing examples of attachment devices (figure by C. Tsoraki).

However, two of the ‘valkyrie’ figures (O5 & O6) show facets in the gap between the feet, suggesting that the artefacts may have been fastened/suspended additionally/alternatively at the ‘bottom’. Could these miniatures have been suspended upside down? Moreover, the loops of three objects were broken (O1, O5 & O9), though whether broken during use or after deposition is difficult to determine. Despite the break, O9 displays perpendicular striations through the loop on its reverse side. These striations do not seem to indicate suspension from a necklace, but instead give the impression of tight binding/wrapping against another object or material. Likewise, two objects display partial flattening of their attachment loops, suggesting that the figures were attached to hard materials, rather than being suspended against softer materials such as fabric or human flesh. A ubiquitous interpretation as objects of bodily adornment does not, therefore, fit with our results, raising the question of what else the figures may have been attached to—other metal objects, wooden structures (posts?) or other materials entirely (see below).

Traces of object circulation were also considered, again finding a high degree of diversity. Two of the ‘valkyries’ (O6 & O9) show substantial rounding of edges suggesting extensive handling over time. Long-term object curation in the Viking Age is apparent, for example, in the Aska burial, which contained O1 and O2 among an artefact assemblage spanning centuries of production (Arwill-Nordbladh Reference Arwill-Nordbladh, Back Danielsson and Thedéen2012). The potential for multigenerational curation opens further horizons for the intra-actions miniature bodies may have seen, of long-term or interweaving storage, display and wear. The rounded edges of O6 and O9 may stem, for example, from fabric wrapped around the surface of the figures or from their storage in pouches, challenging the idea that their primary function was display. Wrapping or ‘dressing’ anthropomorphic imagery in fabric is a specific form of intra-acting with miniature bodies that finds parallels in broadly contemporaneous gold-foil figures (Back Danielsson Reference Back Danielsson, Alberti, Jones and Pollard2013) and the so-called Buddha from Helgö (Gyllensvär Reference Gyllensvärd, Clarke and Lamm2004: fig. 4). These examples testify to the dressing of anthropomorphic bodies through the addition of gold props, such as belts and necklaces, and leather arm- and neck-rings, respectively. While the exact nature is difficult to pinpoint, it seems clear that some artefacts demanded regular engagement.

The miniscule size of many of the artefacts also casts doubt on whether visual engagement was their primary purpose (cf. Bailey Reference Bailey and Dent2014; Weismantel Reference Weismantel and Watts2014; Eriksen Reference Eriksen2022). How visible, for example, would the Rällinge spirals be in a world without cameras and microscopes; could they only be experienced under certain conditions (e.g. lit at a specific angle, or through touch)? Microwear on the chin of the Aska head (O1) may hint at repeated rubbing or handling (and the importance of touch), though no such traces were found on the Aska pregnant body (O2), despite previous speculation (Arwill-Nordblahd Reference Arwill-Nordbladh, Bergerbrant and Sabatini2013b: 415).

In contrast to the well-rounded presentation and potential long-term curation of O6 and O9, O4 and O5 present crisp edges and were thus used sparingly or even made shortly before deposition. Might they have been made for deposition, rather than for use as bodily adornment? Perhaps some objects were neither to be suspended from a body or object, nor manipulated through wrapping or handling—perhaps they were always meant to accompany the dead.

Breaking and mending bodies

Two of the miniature bodies have undergone further forms of manipulation in terms of breakage. The Rällinge figure (O3) is missing part of its left arm. Though we cannot say whether this break was accidental or intentional, the upper fractured end of the arm has a bevel near the missing part, suggesting that the place of breakage was reworked (Figure 5). Likewise, the lower fractured end is smoothed but this may be affected by patination/corrosion. However the breaks occurred, care was taken to rework the upper (and perhaps also the lower) fracture and to smooth the stumps. Furthermore, the choice not to mend or replace the broken arm, but to smooth the place of the fracture instead, seems deliberate.

Figure 5. Traces of reworking on the Rällinge figurine (figure by C. Tsoraki; Rällinge photograph by O. Myrin, Swedish Historical Museums, CC BY 4.0).

The second object that has seen further manipulation is the Aska head (O1) interpreted as the mount of a völva staff (Price Reference Price2019: 104). However, it seems the head was originally attached to a body—it was not cast as a detached head from the beginning. Strikingly, here we found evidence for active percussion at the base of the head (Figure 6), meaning that at some stage of the object’s history, it was actively and intentionally decapitated and thereby split into two objects.

Figure 6. Microwear traces on the Aska head: A–B) percussive traces at the base of the head; C) smooth finishing on the interior ridge of the chin (figure by C. Tsoraki; Aska photograph by O. Myrin, Swedish Historical Museums, CC BY 4.0).

What happened to the rest of the metal body, and how it may have originally looked, we do not know. In contrast to the Rällinge figure, little care seems to have been taken to smooth the edges after breaking the object. Perhaps it did not matter, or perhaps its decapitation was an important part of the object biography, rendering the traces of this action visible.

Deposition

Ultimately, at some point the decision was made to take these objects out of circulation, moving them from whatever prior relationships they were engaged with to new locations and situations. The majority (six out of 10) were deposited in burials—that is, they were to be associated with human remains, as well as an assemblage of artefacts. The others were deposited in hoards or are stray finds.

The practice of linking anthropomorphic objects with dead bodies through burials is little remarked upon in the literature as the objects carry an a priori association with personal adornment reflecting ‘wealth’ or ‘(ritual) status’. Yet, selecting artefacts to accompany the dead rather than circulating them as heirlooms or through systems of gift exchange is a highly intentional act. Why, at this moment, was the object taken out of the world of the living? Why did it belong with this specific dead body? What kinds of engagements were changing, cut off or made new at the moment of deposition? While we may never be able to answer these questions in full, moving from objects as static ‘symbols’ of deities or status, and thinking more deeply about the social fabric they were part of, centres them in practice and in affective relations rather than as passive objects of ‘art’ or ‘ritual’.

The placement of artefacts in burials may speak to their intra-action. The O5 ‘valkyrie’ was found in the midsection of a burial, alongside beads, a cross and a miniature chair and shield. Original documentation indicates that it was found attached to another silver artefact, interpreted as a silver reliquary (Arbman Reference Arbman1940: 298–300). While the publication notes this attachment as likely accidental, we see little reason to outright dismiss the original observation (Figure 7). If this was the original mode of deposition, it raises two interesting observations: first, we may here have an example of how some ‘amulets’ could be attached to hard surfaces/other objects (possibly upside down); second, the combination of Christian and pre-Christian artefacts provides further richness and nuance in our understanding of how these objects may have been worlded.

Figure 7. Location of O5 in the burial at Birka, and an interpretation of the object attached to a reliquary (figure by M.H. Eriksen; grave drawing from Arbman Reference Arbman1940: 299).

The placement of O5 in the burial does support its use for bodily adornment; however, three objects were found in cremation burials. In these cases, the metal bodies cannot have been burnt on the pyre with the human body as an object of personal adornment, as the silver would have melted. Rather, the artefacts—if indeed used by the dead—would necessarily have been removed from the body before burning and reunited with it after, and were thus intentionally preserved intact, in contrast to the human remains transformed through fire.

Finally, we explore the a priori assumption that the ‘valkyrie’ figures were uniquely connected to women’s bodies. None of the burials that contained the objects examined here include skeletal remains securely sexed as female; while two have been gendered female based on the assessment of grave-goods (which comes with its own biases). Meanwhile, O7—a gilded ‘valkyrie’ figurine—was found with human remains osteologically and genetically sexed as male (Arrhenius Reference Arrhenius1990; Rodríguez-Varela et al. Reference Rodríguez-Varela2023: e4), challenging the assumption that these objects are exclusively connected with women’s bodies. Even when biological sex can be genetically or osteologically estimated, we cannot assume that this was the causal connection linking the human and object bodies, nor is—famously—deposition in death the same as use in life. Linked assumptions about these objects seem to steer interpretations. Thus, taking a step back and allowing the objects to emerge, rather than forcing them into assumed categories, allows us to see them in a new light.